Geochimica et Cosmochimica Acta, Vol. 68, No. 20, pp. 4023-4038, 2004
Copyright © 2004 Elsevier Ltd
Printed in the USA. All rights reserved
0016-7037/04 $30.00 ⫹ .00
Pergamon
doi:10.1016/j.gca.2004.04.009
Long term atmospheric deposition as the source of nitrate and other salts in the Atacama
Desert, Chile: New evidence from mass-independent oxygen isotopic compositions
GREG MICHALSKI,1,* J. K. BÖHLKE2 and MARK THIEMENS1
1
Department of Chemistry & Biochemistry, University of California San Diego, La Jolla, CA 92093-0356, USA
2
United States Geological Survey, 431 National Center, Reston, VA 20192, USA
(Received August 19, 2003; accepted in revised form April 7, 2004)
Abstract—Isotopic analysis of nitrate and sulfate minerals from the nitrate ore fields of the Atacama Desert
in northern Chile has shown anomalous 17O enrichments in both minerals. ⌬17O values of 14 –21 ‰ in nitrate
and 0.4 to 4 ‰ in sulfate are the most positive found in terrestrial minerals to date. Modeling of atmospheric
processes indicates that the ⌬17O signatures are the result of photochemical reactions in the troposphere and
stratosphere. We conclude that the bulk of the nitrate, sulfate and other soluble salts in some parts of the
Atacama Desert must be the result of atmospheric deposition of particles produced by gas to particle
conversion, with minor but varying amounts from sea spray and local terrestrial sources. Flux calculations
indicate that the major salt deposits could have accumulated from atmospheric deposition in a period of
200,000 to 2.0 M years during hyper-arid conditions similar to those currently found in the Atacama Desert.
Correlations between ⌬17O and ␦18O in nitrate salts from the Atacama Desert and Mojave Desert, California,
indicate varying fractions of microbial and photochemical end-member sources. The photochemical nitrate
isotope signature is well preserved in the driest surficial environments that are almost lifeless, whereas the
microbial nitrate isotope signature becomes dominant rapidly with increasing moisture, biologic activity, and
nitrogen cycling. These isotopic signatures have important implications for paleoclimate, astrobiology, and N
cycling studies. Copyright © 2004 Elsevier Ltd
spheric deposition was a significant source of nitrate in the
Atacama Desert, with potential inputs including sea spray,
ammonia deposition and nitric acid deposition (Claridge and
Campbell 1968) (Ericksen 1979) (Ericksen 1981). The atmospheric deposition source is strongly supported by isotopic
measurements indicating that the Atacama Desert nitrate has
␦18O values higher than that of atmospheric O2 (Böhlke et al.
1997), similar to modern rainwater nitrate (Kendall 1998).
However, quantifying the influence of atmospheric deposition
from these measurements is difficult because of uncertainties
regarding how nitrate ␦18O values are affected by nitrification
or denitrification in hyper-arid environments and how modern
atmospheric nitrate isotopes are affected by anthroprogenic
NOx emissions. Most authors agree that the nitrate and other
soluble salts are most likely the result of multiple sources, but
the relative magnitudes of the different sources remain uncertain more than one hundred years after Darwin’s publication.
In addition to nitrate, the nitrate ore deposits of the Atacama
Desert also contain substantial amounts of chloride and sulfate
(Ericksen 1981). Composite analyses of ore material indicate
that these major anions were present in subequal molar
amounts, but sulfate and chloride also are more widespread in
the region, and there is considerable heterogeneity within the
ore deposits because of selective remobilization and re-crystallization. For example, calcium sulfate minerals typically are
more abundant near the land surface because of their relatively
low solubility in comparison to nitrate and chloride minerals,
which are found at greater depths (Ericksen 1983; Chong 1994;
Rech et al. 2003). The origin of the sulfate minerals has been
attributed to multiple sulfate sources including sea salt deposition near the coast and rock weathering in regions further
inland (Rech et al. 2003). Volcanism has also been suggested as
1. INTRODUCTION
The Atacama Desert of northern Chile is one of the driest
environments on Earth, with annual rainfall averaging less than
1–2 mm, and it is the premier source of natural nitrate minerals.
Although minor amounts of nitrate are present in other desert
environments, the nitrate ore deposits of the Atacama Desert
are unique because of their high nitrate concentrations, and they
have been mined for export as natural fertilizer since ⬃1830
(Ericksen 1983). Theories on the origin of the Atacama Desert
nitrate deposits date to Darwin, who suggested that they formed
at the inland margin of an oceanic extension (Darwin 1871).
Subsequent studies have suggested a wide variety of sources
including nitrification of ammonia derived from seaweed, vegetation in saline lakes, or bird guano (see Ericksen and
references therein) (Ericksen 1983). The nitrate also has been
attributed to nitrogen fixation and nitrification by microorganisms in the desert soils or in distant Andean salars (dry lakes),
from which it may be transported by wind or groundwater to
the Atacama basin (Ericksen 1983). This hypothesis is supported by the occurrence of nitrogen fixing bacteria associated
with high concentrations of nitrate in playa sediments in the
Mojave Desert and great basin region of the southwestern
United States (Leatham et al. 1983). Others have argued that
the nitrate is the result of magmatic processes that generate
nitrate through unidentified catalytic reaction (s) involving water and magmatic nitrogen gases (Chong 1994). Subsequent
erosion of the igneous rock by hydrothermal activity and transport of the soluble ions is invoked to explain the regional
placement of the nitrate. Several authors postulated that atmo-
* Author to whom correspondence
([email protected]).
should
be
addressed
4023
4024
G. Michalski, J. K. Böhlke, and M. Thiemens
a major constituent of the sulfate budget in the region (Ericksen
1981).
The Atacama Desert is located in northern Chile (between
20°S and 26° S) in the rain shadows of the Andes Mountains to
the east and the Coastal Cordillera (1000 –3000m) to the west
(Fig. 1). The local distributions of the salt deposits are believed
to be largely a function of remobilization of existing salts by
occasional rainfall, eolian transport (Berger and Cooke 1997),
capillary action (Mueller 1968), groundwater discharge, possibly hydrothermal reworking (Pueyo et al.1998), and sedimentation. The spatial distribution, mineralogy, and petrology of
the salt deposits in the region have been described previously in
relation to these transport mechanisms (Ericksen 1981; Searl
and Rankin 1993; Pueyo et al. 1998).
The current paper reports the use of oxygen isotope measurements, specifically ⌬17O values, as a tool to probe the
influence of atmospheric deposition (wet and dry) as a source
of salts in the Atacama Desert region. In addition, it provides
quantitative analyses of the atmospheric chemistry and deposition dynamics involved in nitrate and sulfate accumulation.
The current focus is on the use of stable isotopes 16O, 17O and
18
O to determine the origin of the oxyanions themselves, rather
than their subsequent redistribution and mineralogy.
1.2. Theory of Isotope Distribution and Mass
Independent Fractionations
Stable isotope theory has been used successfully in geochemistry for nearly 50 yr, but it is only in the last few years
that the use of mass-independent isotopic compositions has
yielded new information about terrestrial systems that was
unattainable using traditional isotopic methods (Thiemens
1999). In most geochemical transformations, elements partition
their isotopes into different phases during kinetic and equilibrium reactions with this partitioning (fractionation) being dependent on temperature and the relative mass differences of the
isotopes. These fractionations are quantified by measuring the
isotopic ratios of the sample relative to an accepted standard
and are reported in delta notation (in units of per mil):
␦% ⫽
冉
RSample
RStd
冊
⫺ 1 ⫻ 1000
(1)
where R is the ratio of the minor to major isotopes (e.g.,
18
O/16O or 17O/16O). For elements that have three stable isotopes, the mass dependence is extended to both minor isotopes and is determined by relative mass differences of the two
minor isotopes relative to the third (major) isotope, which for O
is ⬃ 1/2. This relationship is quantified by ␣17/16 ⫽ (␣18/16)0.52,
where alpha is the fractionation factor for a given reaction
(Matsuhisa et al. 1978; Young et al. 2002) and can be converted
to delta notation as approximately ␦17O ⬇ 0.52⫻␦18O (Thiemens 1999; Miller, 2002). Deviations from this rule are quantified by capital delta notation
⌬17O ⬇ ␦17O ⫺ 0.52 ⫻ ␦18O
Fig. 1. Map of the major nitrate ore fields of the Atacama Desert in
northern Chile with isotope sample localities (modified from Ericksen;
1981).
(2)
Most terrestrial materials have ⌬17O ⬇ 0, including oceanic
and meteoric waters, air O2 (within tenths of a per mil) and
most solids (Miller 2002). The discovery of approximately
equal enrichment of both minor isotopes (␦17O ⬇ ␦18O) in the
formation of ozone (O3 ⌬17O ⬇ 20 –50‰) demonstrated the
exception to this rule (Thiemens and Heidenreich 1983). Initially of interest primarily to atmospheric chemists, photochemically induced mass-independent fractionations have been
found recently in desert varnish (Bao, Michalski, and Thiemens, 2001), sulfate deposits (Bao et al. 2000a), sulfate contained in Antarctica ice cores (Alexander et al. 2002) and
Achaean sulfate and sulfide rocks (Farquhar et al. 2000). Minerals containing NO3⫺ and SO4⫺2 are of particular interest
because they are stable compounds that do not undergo isotopic
exchange with water (Spindel 1969), which allows the preservation of their isotopic signatures in ice, water, soils, and the
geologic record. The ⌬17O (and ⌬33S) signatures in these
substances provide a conservative tracer to past photochemical
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
processes and a measure of the magnitude of atmospheric
inputs to various geologic and hydrologic environments.
1.2.1. Origin of ⌬17O values in photochemically produced
NO3⫺ and SO4⫺2
Several accounts on the origin of the salts in the Atacama
Desert have invoked “atmospheric or electrochemical N2 fixation” but provided few details on the atmospheric chemistry or
deposition dynamics. It is therefore important to examine the
photochemical mechanisms that produce NO3⫺ and SO4⫺2 and
how different reaction pathways can mediate the O isotopic
composition of these substances.
One of the most important atmospheric oxidants and a key
reactant in the formation of NO3⫺ and SO4⫺2 in the atmosphere is O3 (Seinfeld and Pandis 1998). Numerous experimental, theoretical and in situ measurement efforts (Thiemens
2001) have demonstrated that O3 can have ⌬17O values from
20 to 50‰. It is believed the origin of these ⌬17O enrichments
is the result of differential lifetimes of excited O3 molecules
depending on whether the symmetries of the molecules have
been altered by isotopic substitution (Hathorn and Marcus
1999; Gao and Marcus 2001). The ⌬17O values generated
during O3 formation are well constrained for various temperature and pressure regimes (Morton et al.1990). ⌬17O enrichments have also been detected recently in aerosol and rainwater
NO3⫺ (20 –30‰) (Michalski et al. 2003) and SO4⫺2 (0.2–2‰)
(Lee 2000). Currently, the accepted mechanism for producing
⌬17O in both NO3⫺ and SO4⫺2 is through mass transfer of O
atoms during gas- and aqueous-phase oxidation reactions involving O3. Laboratory experiments (Savarino et al. 2000) have
documented this mass transfer scheme in aqueous phase oxidation of SO2, and the ⌬17O values measured in atmospheric
NO3⫺ samples have been simulated using an isotopic-photochemical box model (Michalski et al. 2003).
⫺
important in urban regions and those involving dimethylsulfide
(DMS) are more important in the marine boundary layer (Yvon
et al. 1996). Since O atoms are conserved in each HNO3
reaction pathway the ⌬17O of HNO3 can be determined by the
mass balance equation:
⌬17O(HNO3) ⫽ ␩(⌬17O(3)) ⫹ ␤(⌬17O(4)) ⫹ ␹(⌬17O(5))
(6)
Where ␩, ␤, and ␹ are the fractions of HNO3 produced by
reaction pathways 3, 4 and 5 and (⌬17O(3,4,5)) is the ⌬17O
value of HNO3 generated by each formation pathway. To
evaluate (⌬17O(3,4,5)) one needs measured or calculated ⌬17O
values for the reactants (O3, NO2, OH and H2O). The ⌬17O
values for H2O have been measured and found to equal zero for
all natural terrestrial waters (Meijer and Li 1998). In addition,
the rapid isotopic exchange between H2O(g) and OH relative to
the sink reaction (3) (Dubey et al. 1997) dictates that ⌬17O⬃0
for OH in the troposphere (Lyons 2001). Experimental pressure- and temperature-dependent rate data were used to determine O3 ⌬17O values, which are ⬃35‰ for tropospheric O3 at
mid latitudes (appendix I). Experimental data were used for this
estimate because possible sampling artifacts have cast doubt on
the accuracy of published tropospheric O3 ⌬17O in situ measurements (Krankowsky et al. 1995; Johnston and Thiemens
1997). The ⌬17O value of tropospheric NO2 is not known but
can be estimated based on its formation mechanism. NO2 is
generated by the oxidation of NO by either O3, HO2 or peroxy
radicals (ROx). NO2 is also readily photolyzed in visible/uv
light to give back NO and O3 in the classic NOx cycle:
NO ⫹ O3 → NO2 ⫹ O2
(7)
NO ⫹ · HO2(ROx) → NO2 ⫹ · OH 共ROx)
(8)
共 3P 兲
(9)
O(3P) ⫹ O2 ⫹ M → O3 ⫹ M
(10)
NO2 ⫹ hv → NO ⫹ O
1.2.2. Photochemical model for ⌬17O in NO3⫺atm
⫺
Atmospheric nitrate (NO3 atm ⫽ HNO3(gas) ⫹ NO3 salts)
is produced by photochemical reactions involving NOx that
generate gas phase nitric acid. Nitric acid (HNO3) can then
react with sea salt or crustal aerosols in neutralization reactions
that generate particulate NO3. Gas phase HNO3 is also highly
reactive on land surfaces (dry deposition), and both HNO3 and
NO3⫺ aerosols are highly soluble (wet deposition), so that wet
and dry deposition are the primary mechanisms for removal of
NO3⫺atm from the atmosphere.
There are three major reaction pathways involving NO2, O3,
OH and H2O that produce HNO3 with varying ⌬17O values
(Michalski et al. 2003):
4025
NO2 ⫹ · OH ⫹ M → HNO3 ⫹ M
(3)
NO2 ⫹ O3 → NO3 ⫹ O2
(4)
Because the NOx photo-stationary state (7–10) achieves steady
state 3 orders of magnitude faster than the sink reactions (3–5)
(Seinfeld and Pandis 1998) the O atoms in NO2 achieve isotopic equilibrium with O3 and HO2 (ROx). The ⌬17O value of
HO2 is assumed to be zero, which simplifies the problem and is
justified because measured values of ⌬17O in H2O2 (the main
product of HO2 self reaction) are ⬃1.4‰ (Savarino and Thiemens 1999). Because the fraction of NO2 oxidized by peroxy
radicals on a global average is only approximately 10%, this
assumption only introduces errors at the 0.1‰ level. The resulting NO2 ⌬17O values are approximately 31.5‰. Using Eqn.
3– 6 with O3 ⌬17O ⫽ 35‰ and both H2O and OH ⌬17O ⫽ 0,
we obtain:
NO3 ⫹ HC, DMS → HNO3 ⫹ products
(4b)
⌬17O(HNO3) ⫽ ␩(21 ‰ ) ⫹ ␤(33 ‰ ) ⫹ ␹(27 ‰ )
NO2 ⫹ NO3 ↔ N2O5
(5)
N2O5 ⫹ H2O ⫹ surf共aq兲 → 2 HNO3
(5b)
The branching ratios (␩, ␤, ␹) are determined using either
simple photochemical box models or more complex global
climate models with detailed atmospheric chemistry subroutines (see below).
The NO3 radical reactions involving hydrocarbons (HC) are
(11)
4026
G. Michalski, J. K. Böhlke, and M. Thiemens
1.2.3. Photochemical model for ⌬17O in nss-SO4⫺2
Sulfate aerosols in the atmosphere have two main sources,
sea spray and gas-to-particle conversion of SO2, which is
referred to as non-sea-salt SO4⫺2 (nss-SO4⫺2). The nss-SO4⫺2
is produced by both gas phase (homogenous) and aqueous
phase (heterogeneous/cloud processing) SO2 oxidation (Seinfeld and Pandis 1998). Sulfate ⌬17O values are generated
during the aqueous phase oxidation of SO2 to SO4⫺2 (Savarino
et al. 2000), which is initiated by equilibrium between gas
phase SO2 and liquid water droplets:
SO2 ⫹ H2O共1兲 ↔ H2SO3共aq兲 ↔ H⫹ ⫹ HSO3⫺ ↔ H⫹ ⫹ SO3⫺2
(12)
All S species in (12) have the same oxidation state and are
collectively termed S(IV). Isotopic exchange between O in the
S(IV) species and H2O is very fast and rapidly achieves isotopic equilibrium with tropospheric H2O (Newman, Krouse and
Grinenko 1991), for which ⌬17O ⫽ 0. Because the oxidation of
reduced S species such as H2S and DMS proceeds though
S(IV), isotope effects during the initial oxidation of reduced S
do not affect the final ⌬17O value of nss-SO4⫺2 and are not
explicitly included in the model. The aqueous-phase oxidation
of S(IV) to SO4⫺2 is carried out primarily by O3 and H2O2,
with the branching ratio depending on the localized concentration of each reactant and the pH of the water droplet (Lin and
Chameides 1991):
S(IV) ⫹ O3 → SO4⫺2 ⫹ O2
(13)
S(IV) ⫹ H2O2 → SO4⫺2 ⫹ H2O
(14)
The nss-SO4⫺2 ⌬17O values depend on the branching ratio of
reactions 13 and 14 and on the ⌬17O values of O3 (35‰) and
H2O2 (⬃1.4‰) (Savarino and Thiemens 1999). Experimental
studies have shown that one O atom is transferred to S(IV)
during oxidation by O3 (yielding SO4⫺2 ⌬17O ⬃ 9‰), that two
O atoms are transferred during H2O2 oxidation (yielding
SO4⫺2 ⌬17O ⬃ 0.7‰) and that the reaction dynamics generate
no mass-independent isotope effects (Savarino et al. 2000). No
⌬17O signal arises from gas phase SO2 oxidation because all
three sources of O atoms in the gas phase oxidation scheme
(OH, O2, H2O, and SO2) have ⌬17O ⬃ 0 and the reaction itself
has been shown to proceed via a conventional mass-dependent
fractionation pathway (Savarino et al. 2000):
SO2 ⫹ OH → HSO3
(15)
HSO3 ⫹ O2 → SO3 ⫹ HO2
(16)
SO3 ⫹ H2O → H2SO4 (⌬17O ⫽ 0)
(17)
Gas phase H2SO4 is highly hygroscopic and quickly nucleates
with H2O to form nanometer-size droplets of sulfuric acid
(H2SO4). The ⌬17O of nss-SO4⫺2 is thus given by:
⌬17O nss ⫺ SO4⫺2 ⫽ ␧[(␾ ⫻ 9 ‰ ) ⫹ (␥ ⫻ 0.7 ‰ )] (18)
where ␾ and ␥ are the oxidation branching ratios for H2O2 and
O3 respectively (␾ ⫹ ␥ ⫽ 1) and ␧ is the fraction of SO2
oxidized in the aqueous phase. Computer modeling is required
to evaluate these parameters. In fact, SO4⫺2 ⌬17O measure-
ments can be used to test whether computer models are correctly partitioning the relative importance of homogenous versus heterogeneous SO2 oxidation (see below).
Aerosols originating from dried seawater droplets generated
by bursting air bubbles on the sea surface contain ions in
approximately the same ratios as seawater ions (molar Na⫹/
Cl⫺ ⫽ 0.86; SO4⫺2/Cl⫺ ⫽ 0.052). This is a significant source
of SO4⫺2 aerosols, as the global atmospheric flux of sea salt is
estimated to be 1.17 ⫻ 1016 g yr⫺1. (Gong et al. 1997). The O
isotope ratios of oceanic SO4⫺2 have been measured and have
⌬17O ⫽ 0 (Lee and Thiemens 2001), as expected since oceanic
SO4⫺2 is derived mainly from mass-dependent kinetic oxidation of organic and sulfide S phases. The ⌬17O in aerosol
SO4⫺2 is therefore a function of the relative contribution of sea
salt versus nss-SO4⫺2 production, of the amount of homogeneous versus heterogeneous gas phase SO2 reactions, and of the
relative importance of O3 versus H2O2 in aqueous phase oxidation. If Rnss/ss is the ratio of nss-SO4⫺2 to sea salt SO4⫺2,
then
⌬17O SO4⫺2 ⫽ Rnss/ss ⫻ ⌬17O nss ⫺ SO4⫺2 ⫽ Rnss/ss
⫻ ␧[(␾ ⫻ 9 ‰ ) ⫹ (␥ ⫻ 0.7 ‰ )]
(19)
2. METHODS
We have performed isotopic measurements to detect ⌬17O
anomalies in NO3⫺ and SO4⫺2 salts from the Atacama Desert
and Mojave Desert that were described previously by Böhlke et
al. (1997). More complete descriptions of the mineralogy and
locations of the individual samples can be found in Ericksen
(1981), Ericksen et al. (1988), and Böhlke et al. (1997), with
summary descriptions provided in Table 1. The Atacama
Desert samples are all from the nitrate ore fields and are
referred to as caliche-type deposits (Ericksen, 1981). Most of
the samples are salt crystals or mixed salt cements in soil
samples collected from open pits in the nitrate mines. Two of
the samples (375 and USGS35) were commercial nitrate salts
that were purified by fractional crystallization. The Mojave
Desert salt samples are from caliche-type deposits in the clayhills region near Death Valley, California (Ericksen et al.,
1988). Additional NO3⫺ samples were extracted from ground
water pumped from the Irwin basin in the Mojave Desert of
California. Ground-water NO3⫺ in the Irwin basin is believed
to be derived largely from leaching and recharge of NO3⫺
formed naturally within the unsaturated zone, with some localized occurrences of anthropogenic NO3⫺ resulting from wastewater disposal (Densmore and Böhlke, 2000).
For ⌬17O analysis, archived NO3⫺ salts (as KNO3) were
converted to AgNO3 using a cation exchange membrane (Dionex AMMS-4ml suppression column) utilizing 0.002M
Ag2SO4 as the regenerant. The AgNO3 solutions were then
freeze-dried in silver capsules (99.9% pure) and thermally
decomposed into O2 and NO2 by heating in vacuo at 520°C.
The NO2 gas was cryogenically removed using a series of
liquid nitrogen traps and the O2 was collected for isotopic
analysis. A Finnigan-MAT 251 isotope ratio mass spectrometer
was used to determine 18O/16O and 17O/16O isotope ratios in
the O2, from which ⌬17O values were determined with a
precision of ⫾0.2 ‰. Complete details of the experimental
method may be found in Michalski et al. (2002). In addition to
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
4027
Table 1. Isotopic and chemical data for nitrate and sulfate from the Atacama Desert and Mojave Desert.
Sample
⌬17O ␦18O ⌬17O ␦18O ␦15N ␦34S
NO3⫺ NO3⫺ SO4⫺2 SO4⫺2 NO3⫺ SO4⫺2
Cl/N
mol.
S/N
mol.
b
0.73
0.001
b
0.001
0.000
Long.
Lat.
Description
b
69° 55= W
69° 59= W
b
19° 40= S
19° 39= S
b
Of. Dolores; blocky NaNO3 crystals
Of. San Antonio; blocky NaNO3
crystals
Of. San Antonio; blue granular
mixed salts
Of. San Antonio; acicular NaNO3
crystals
Of. San Pablo; acicular NaNO3
crystals
Pintados; granular mixed salts
Of. Victoria; darapskite crystals
Of. Victoria; granular mixed salts
Of. Santa Fe; granular mixed salts
Of. Maria Elena; fertilizer product,
NaNO3 pellets
Of. Pedro de Valdivia; granular
mixed salts
Pampa Yumbas; salt cemented soil
Of. Alemania; humberstonite-rich
salt
Of. Lautaro; blocky NaNO3 crystals
Of. Lautaro; acicular NaNO3
crystals
Of. Santa Luisa; mixed salts
fertilizer product, granular NaNO3
Atacama
Desert Salts
439
443
a
18.0
15.2
a
55.1
50.1
a
na
na
a
na
na
b
⫺1.0
⫺1.0
444
17.2
44.8
0.9
2.9
⫺4.9 ⫺2.8
0.31
0.026
69° 59= W
19° 39= S
447
15.6
45.8
na
na
⫺2.2
na
1.11
0.000
69° 59= W
19°39= S
461
19.6
60.2
na
na
⫺1.0
na
0.95
0.012
69° 42= W
20° 24= S
458
91c-01
91c-02
87c-20
375
16.2
16.6
16.4
13.7
19.2
49.1
52.8
51.9
44.4
52.9
0.7
0.4
na
2.4
na
2.7
5.4
na
5.3
na
0.6 ⫺0.2
⫺0.8
na
⫺1.6
na
⫺0.8
2.0
⫺0.4
na
0.56
na
na
0.74
0.014
0.033
na
na
0.047
0.002
69°
69°
69°
69°
69°
20°
20°
20°
21°
22°
8304
18.7
52.1
na
na
⫺2.6
na
na
69° 47= W
22° 45= S
1221a
193
19.9
19.4
48.9
50.8
4.0
1.5
5.5
7.3
1.5
0.9
5.4
4.5
0.003
0.013
0.039
2.3
70° 10= W
69° 55= W
24° 55= S
25° 07= S
143
32a
17.6
16.8
52.8
52.4
na
na
na
na
3.0
4.1
na
na
0.017
0.000
0.002
0.001
69° 49= W
69° 49= W
25° 13= S
25° 13= S
205
USGS35
18.0
21.6
52.4
57.5
na
na
4.7
na
2.9
2.7
0.4
na
0.40
0.010
0.017
0.002
70° 08= W
25° 14= S
7.2
21.3
0.6
3.2
4.8 ⫺19.4
6.9
0.93
116° 12= W
35° 42= N
8405c
13.1
35.9
2.0
5.3
0.1
9.9
1.7
0.10
116° 36= W
35° 50= N
8414
12.4
34.5
1.9
5.4
0.0
9.7
1.6
0.080
116° 36= W
35° 50= N
8419
12.6
36.0
0.7
1.8
0.9
5.4 13.6
1.49
116° 28= W
35° 39= N
0.4
0.9
0.4
2.8
2.9
1.4
na
na
na
na
na
na
11.4
7.5
6.7
na
na
na
3.9
8.4
2.3
116° 40= W
116° 41= W
116° 40= W
35° 15= N
35° 16= N
35° 16= N
Mojave Desert
Salts
8401
Mojave Desert
Ground Water
WC1-150
NH1-300
BC1-185
b
na
na
na
8.9
9.4
2.6
46=
43=
43=
34=
37=
W
W
W
W
W
32=
41=
41=
51=
20=
S
S
S
S
S
Bully Hill, Upper Canyon area; salts
leached from soil
Confidence Hills; salts leached from
soil
Confidence Hills; salts leached from
soil
Saratoga hills; salts leached from
soil
Irwin Basin; pumped well
Irwin Basin; pumped well
Irwin Basin; pumped well
a ⫽ this study b ⫽ Böhlke et. al. (1997); Densmore and Böhlke (2000); Böhlke et al. (2003)
the thermal decomposition measurements, we also re-determined the ␦18O values of the KNO3 samples by continuousflow mass spectrometry on CO produced by high-temperature
reaction with carbon (CO-CFIRMS) (Kornexl et al. 1999;
Böhlke et al., 2003). These analyses were calibrated by analyses of reference materials IAEA-NO-3 and USGS35, which
were assumed to have ␦18O values of ⫹25.6 and ⫹57.5 ‰,
respectively (Böhlke et al., 2003). A subset of the original salt
samples consisting of relatively pure NaNO3 crystals was also
analyzed by the CO-CFIRMS method and by the bacterial N2O
method of Casciotti et al. (2002), which yielded similar results.
The new NO3⫺ ␦18O values determined by CO-CFIRMS and
the bacterial method are systematically higher than the values
determined previously by an off-line preparation technique
(Böhlke et al. 1997; Revesz et al. 1997). The systematic difference between these two sets of results is consistent with
reported ␦18O scale contraction in samples prepared by off-line
combustion in sealed glass tubes (Revesz and Böhlke 2002).
The NO3⫺ ␦18O values reported in this paper are averages of all
available normalized data produced by the CO-CFIRMS and
bacterial methods, which are believed to be more reliable than
the previously published off-line values and have a reproducibility of approximately ⫾0.3 to 1.0 ‰ (1-␴).
Sulfate samples for ⌬17O analysis were obtained by hydrating bulk salt samples with Millipore water, heating the solution
to 80°C and slightly acidifying the solution using HCl. Sulfate
was then precipitated from solution by adding excess BaCl2.
The resulting BaSO4 solid was concentrated by centrifugation
(10 min) followed by decantation of the supernatant. The
BaSO4 pellets were rinsed in 10 mL of Millipore water, recentrifuged and decanted, and dried at 100°C overnight. O2 was
generated by fluorination of BaSO4 (⬃8mg) using BrF5 gas and
a CO2 IR laser as described by Bao and Thiemens (2000b). The
␦18O values were determined by the CO-CFIRMS method in
4028
G. Michalski, J. K. Böhlke, and M. Thiemens
BaSO4 precipitates prepared previously and calibrated against
analyses of NBS-127 BaSO4, for which a value of ⫹8.6 ‰ was
assumed (Kornexl et al., 1999; Böhlke et al., 2003). Experimental precision for sulfate was ⫾0.2 ‰ for ⌬17O and ⫾0.5 ‰
for ␦18O, based on replicate samples. The ␦18O data (NO3⫺ and
SO4⫺2) presented here are from the CO-CFIRMS and bacterial
methods because the thermal decomposition and laser fluorination techniques used for ⌬17O measurements can have massdependent fractionation variations of 1– 4‰.
3. RESULTS OF ANALYSES
The ⌬17O and ␦18O values for both NO3⫺ and SO4⫺2
minerals are given in Table 1, along with the ␦15N and ␦34S
values given by Böhlke et al. (1997). The data are also shown
in three-isotope plots in Figures 2 and 3. The NO3⫺ samples
from the Atacama Desert have large ⌬17O values (14 –21‰),
which are a factor of 5 larger than those of any other terrestrial
minerals reported to date (Thiemens et al. 2001). The highest
⌬17O value (21.5‰) is at the lower end of ⌬17O values measured in aerosol and precipitation NO3⫺ in a variety of environments (21–31‰) (Michalski et al. 2003 Michalski unpublished data). Smaller though significant ⌬17O values are also
observed in SO4⫺2 that is associated with the Atacama Desert
nitrate ore deposits (0.4 – 4 ‰). These values fall within the
high end of the range of ⌬17O values measured in aerosol
SO4⫺2 in the northern hemisphere (Lee 2001; Bao, Michalski,
and Thiemens 2001) and in some SO4⫺2 deposits associated
with ancient volcanic eruptions (Bao et al. 2000a). Since the
only processes known to cause ⌬17O values of these magnitudes on Earth are photochemical reactions in the troposphere
and stratosphere, the new data provide strong evidence that
photochemical production and subsequent deposition is responsible for a substantial portion of both the NO3⫺ and SO4⫺2 in
minerals found in the nitrate ore deposits of the Atacama
Desert.
Among the NO3⫺ samples, there is a positive correlation
between ⌬17O and ␦18O (Fig. 2b), but no apparent correlation
between ⌬17O and either mineral type or ␦15N (Table 1). The
⌬17O and ␦18O values are highest in NO3⫺ from the Atacama
Desert nitrate ore deposits and somewhat lower in surficial
nitrate mineral samples from the Death Valley region of the
Mojave Desert, California. Samples of NO3⫺ in ground water
from the Mojave Desert have ⌬17O and ␦18O values near 0, as
expected for NO3⫺ formed by biologic processes in moist
environments. Similarly, among the SO4⫺2 samples, there appears to be a positive correlation between ⌬17O and ␦18O, with
values slightly higher on average in the Atacama Desert samples. The lowest SO4⫺2 ⌬17O values are near 0, as expected for
either sea-salt SO4⫺2 or SO4⫺2 formed by chemical or microbial S oxidation in soils and most other terrestrial aqueous
environments.
It is important to recognize that the correlations observed in
the three oxygen isotope plots can have a variety of meanings.
For example, plots of ␦17O/␦18O of O3 formed in laboratory
experiments that remove O3 immediately after formation follow a slope of ⬃1 (Thiemens and Heidenreich 1983), yet
stratospheric O3 samples follow a slope of ⬃0.62 (Lammerzahl
et al. 2002). This can be explained by mass-dependent fractionations, such as photolysis and catalytic destruction that
Fig. 2. Nitrate O isotopic variations in samples of Atacama Desert
caliche-type salts, Mojave Desert caliche-type salts, and Mojave Desert
(Irwin Basin) groundwater nitrate: a) ␦17O versus ␦18O; b) ⌬17O versus
␦18O. Data shown for comparison include atmospheric nitrate in rain
and aerosols from California and Antarctica (Michalski et al., 2003
unpublished data) and nitrate isotopic reference materials IAEA-N3
and USGS35 (Michalski et al., 2002; Böhlke et al., 2003). The terrestrial fractionation line indicates the trajectory followed by mass-dependent isotopic variations. The ozone field is centered on the modeled
value of ⌬17O, with varying mass-dependent fractionation, as explained
in the text. Correlated variations in the NO3⫺atm data are interpreted to
reflect the proportion of O atoms that are derived from O3 or tropospheric H2O, which depends on the photo-oxidation pathway. The fit to
the desert nitrate data is interpreted to indicate varying mixtures of
atmospheric photochemical nitrate (star ⫽ model) and terrestrial microbial nitrate (shaded rectangle at ⌬17O ⫽ 0). The ⌬17O versus ␦18O
plot (b) emphasizes the effects of mass-dependent fractionations (horizontal vectors) that are less obvious in the ␦17O /␦18O plot.
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
Fig. 3. Sulfate O isotopic variations in samples of nitrate-rich caliche-type salts from the Atacama Desert and Mojave Desert. The
terrestrial fractionation line indicates the trajectory followed by massdependent variations. The nss-sulfate field is centered around the modeled value of ⌬17O, with varying mass-dependent fractionation, as
explained in the text. The sulfate isotopic reference material (NBS-127)
is similar isotopically to seawater sulfate. The poorly correlated isotopic variations among the salt samples are consistent with varying
mixtures of atmospheric nss-sulfate, sea-salt sulfate, and terrestrial
microbial sulfate (e.g., from sulfide oxidation; Van Stempvoort and
Krouse, 1994).
follow of slope 0.52, superimposed on the formation line with
slope 1.0. In contrast, the slope of 0.8 observed in NO3⫺atm is
considered to be primarily a mixing line between the two main
isotopic reservoirs contributing to NO⫺atm formation (O3 and
H2O) in the relative absence of fractionation effects (Michalski
et al., 2003). Below it is proposed that the correlations observed
among the desert NO3⫺ samples also represent mixing lines,
but with different end members (NO3⫺atm and microbial nitrification NO3⫺).
4. SOURCES OF NITRATE AND SULFATE IN THE
ATACAMA DESERT
4.1. Isotopic Constraints on NO3ⴚ Sources
To evaluate the proportion of Atacama Desert NO3⫺ attributable to atmospheric photochemistry and deposition we employed the isotope/photochemical box model for NO3⫺atm
⌬17O production presented by Michalski et al. (2003). This
model uses the mass transfer mechanism discussed in the
introduction in conjunction with a photochemical box model
from Yvon et al. (1996) to evaluate ␩, ␤ and ␹ and the
importance of reactions 7 and 8. The input parameters were
representative estimates for trace gas concentrations and emissions in the preindustrial atmosphere in the southern hemisphere given by Holland et al. (1999) and Logan et al. (1981).
Trace gas concentrations were assumed constant through time,
and the temperature was estimated by a simple seasonal tem-
4029
perature model for 20°S. The model yielded an annual mean
NO3⫺atm ⌬17O value of 24‰, slightly less than modern urban
averages (26‰) because of the much smaller concentrations of
NOx and the associated reduced importance of the N2O5 production pathway (5) in comparison to modern polluted regions.
These results are in general agreement with results from global
climate models (GCM) that estimated the proportion of
NO3⫺atm produced by each formation pathway and the corresponding deposition rates (Dentener and Crutzen 1993; Derwent et al. 2003). The GCM of Derwent et al. (2003) yielded ␩
⫽ 0.91, ␤ ⫽ 0, and ␹ ⫽ 0.09, giving a global average HNO3
⌬17O⬃ 24.7‰. Using this basis, the sample-to-sample contribution of NO3⫺atm to the nitrate minerals ranges from 55 to
88%. If each of the samples is assumed to represent an equal
proportion of the total NO3⫺ in the region, the average ⌬17O
value (17.8‰) indicates at least 74% of the NO3⫺ in the
Atacama Desert nitrate ore deposits is the direct result of
atmospheric deposition of photochemically produced NO3⫺.
There appears to be significant local variation in the ⌬17O
values but no systematic trends with longitude or latitude were
observed. The local variation may have at least two interpretations. The first is that all sources of NO3⫺ have constant
fluxes and the ⌬17O variation reflects fluctuations in the ⌬17O
values of NO3⫺atm. The observed seasonal variation in ⌬17O
values of modern NO3⫺atm (⫾5‰) (Michalski et al.2003)
reflects changing oxidation pathways that alter as a function of
solar intensity, NOx concentrations, and temperature. However,
these ⌬17O values should converge on a constant value when
averaged over one or multiple years of collection. Variations in
NO3⫺atm average ⌬17O values over longer time scales are
expected to be small. Only a few per mil difference is observed
(calculated) between ancient and modern NO3⫺atm ⌬17O values and the modern atmosphere has been perturbed by anthroprogenic NOx to a much greater degree than in long term
climatic events. The possibility that tropospheric chemistry
varied spatially and this was reflected in the isotopic composition in the individual deposits is considered unlikely. More
plausible is that NO3⫺ from long-term deposition, with little
integrated ⌬17 O variability, is mixed with NO3⫺ derived from
other mass-dependent sources whose importance varies temporally and spatially. The most likely source of this additional
NO3⫺ is nitrification of NH4⫹ derived from soil organic N (and
possibly marine organic N) that was either co-deposited with
NO3⫺ or produced locally (e.g., by N2-fixing bacteria). Modern
NH4⫹ /NO3⫺ ratios in the remote, clean oceanic atmosphere
are typically 1:1, but may be as high as 2:1, and organic N can
equal NH4⫹ (as N) in coastal regions (Dentener and Crutzen
1994; Peierls and Paerl 1997). Even if some of the deposited
NH4⫹ was volatilized as NH3, only a fraction of the available
N would have to be nitrified to account for the remaining
⬃25% of the mineral nitrate in the absence of local N2 fixation.
This two-source hypothesis is supported by analyses of caliche-type nitrate deposits from the milder, wetter Mojave
Desert, where nitrification is more likely to be an important
source of soil NO3⫺ (Leatham et al.1983). The ␦17O, ⌬17O, and
␦18O values of the Mojave Desert NO3⫺ are presented along
with the Atacama Desert NO3⫺ values in Figure 2. There is a
strong linear correlation between ␦17O and ␦18O values in the
NO3⫺ (R2 ⫽ 0.96), which intercepts the terrestrial mass-dependent fractionation line near the zero intercept. A similar
4030
G. Michalski, J. K. Böhlke, and M. Thiemens
correlation exists between the ⌬17O and ␦18O values. Experiments indicate that nitrate formed by microbial nitrification of
NH4⫹ obtains two of its O atoms from water and one from air
O2 (␦18O ⫽ 23.5‰) (Andersson and Hooper 1983; Hollocher
1984). If this occurs with no associated isotopic fractionation,
then microbial nitrification in the presence of water with an
average ␦18O value of about ⫺6 ‰ or lower, similar to meteoric water values in both regions (Williams and Rodoni 1997;
Aravena et al. 1999), would yield ␦18O values of ⬃4‰ or
lower. Several samples of NO3⫺ from ground water in the
Mojave Desert had ␦18O values of around ⫹1 to ⫹3 ‰ and
⌬17O values of around ⫹0.4 to ⫹0.9 ‰. The NO3⫺ in those
samples has relatively high ␦15N (⫹7 to 11 ‰) and is believed
to have formed biologically by nitrification of both natural and
anthropogenic reduced N (Densmore and Böhlke 2000). Therefore, the linear correlations among ␦18O, ␦17O, and ⌬17O are
interpreted as mixing lines with end members representing
NO3⫺atm and microbial NO3⫺, where the microbial NO3⫺ may
be derived from either atmospherically deposited reduced N or
terrestrial reduced N. The limited extent of mixing indicated by
the Atacama Desert deposits is consistent with as much as
100% of the total N being derived from atmospheric deposition: ⬃75% from direct deposition of NO3⫺atm derived from
photochemical NOx oxidation and ⬃25% from nitrification of
inorganic/organic N that was also deposited from the atmosphere. In contrast, the Mojave Desert caliche NO3⫺ samples
appear to be partially derived from atmospherically deposited
N, but also contain substantial amounts of N from terrestrial N2
fixation, rock weathering, or plant N recycling.
Microbial N2 fixation and nitrification of reduced N in the
Atacama Desert could occur sporadically after occasional rainfall events or during longer periods of wetter climate. NO3⫺
accumulated from nitrification and from NO3⫺atm deposition
may not have mixed completely and this could account for
some of the variability in the measured ⌬17O values of different
samples. For example, long term (⬎100 yr) deposition of
NO3⫺, NH4⫹ and organic N could accumulate these N species
on the desert surface. During rare rain infiltration events, the
NO3⫺ would be leached rapidly to depth, whereas the soil
matrix might retain NH4⫹ and organic N. Subsequent nitrification of the reduced N would produce NO3⫺ that would then
be mixed with NO3⫺atm deposited during the next desiccation/
infiltration cycle and the system could become heterogeneously
mixed depending on water content, which could vary spatially
by topography and soil composition. Partial dissolution and
sequential crystallization during such events would result in
heterogeneous deposits containing zones and veins of different
composition, mineralogy, and O isotopic composition. Numerous cycles of NO3⫺ deposition, nitrification, re-solution, and
re-crystallization would be consistent with complex mineralogic assemblages and sequences described from detailed petrographic studies of the deposits (Searl and Rankin 1993). In
future studies it would be interesting to determine if the nitrate
morphology associated with different dissolution/crystallization dynamics have correlated ⌬17O values.
4.2. Isotopic Constraints on SO4ⴚ2 Sources
To evaluate the contribution of nss-SO4⫺ to total mineral
sulfate we have modeled the possible ⌬17O values for nssSO4⫺2 using various global climate models (GCMs), and the
mass balance considerations discussed in the introduction. The
GCM modeled ⌬17O values (0.4 to 1.2 ‰) are given in Table
2 and are similar to measured values in aerosol and rainwater
SO4⫺2 from La Jolla, California (Lee 2000). However, several
of the Atacama Desert SO4⫺2 ⌬17O values are higher than
those, one by a factor of three. This suggests either that the
GCMs are incorrectly partitioning oxidation pathways or that
the oxidation pathways that generated SO4⫺2 aerosols that
were deposited in the Atacama Desert were markedly different
from those in modern environments. A re-analysis of the GCM
predictions indicates that the former explanation may be more
likely.
GCM’s commonly use simplified aqueous-phase chemistry
to reduce model run times. One of these simplifications is the
“bulk pH” parameterization, in which all cloud droplets are
Table 2. Global Circulation Model estimates of total SO2 oxidation (top) and box model heterogeneous oxidation coupled to OH oxidation from
the GCMs (bottom). ⌬17O for nss-SO4⫺2 are estimated using the mass balance approach with ⌬17O (‰) ⫽ (1.4‰) ⫻ %H2O2 ⫹ (35‰) ⫻ %O3 ⫹
(0‰) ⫻ %OH. GCMs are ECHAM (European Centre Hamburg Model), GISS (Goddard Institute for Space Studies), GOCART (Goddard Global
Ozone Chemistry Aerosol Radiation Transport) NCAR (National Center for Atmospheric Research), MOGUNTIA (Model of the Global Universal
Tracer Transport in the Atmosphere).
Author
GCM Model
% OH
%
O3
% H2O2
nss-SO4 ⌬17O
Feichter (Feichter J. et. al. 1996)
Rasch (Rasch P. J. et. al. 2000)
Barth (Barth M. C. et al. 2000)
Koch (Koch D. et. al. 1999)
Chin (Chin M. et. al. 2000)
Langner (Langner J. and Rodhe H. 1991)
Box Model
O’Dowd (O’Dowd C. D. et al. 2000)
Lin (Lin X. and Chameides W. L.
1991)
Lin
Lin
Lin
Lin
Lin
ECHAM
ECHAM
NCAR
GISS
GOCART
MOGUNTIA
35.9
18
17
29
36
15
13.5
8
0
0
0
na
50.4
74.0
83.0
70.7
64.0
na
1.1
1.2
0.7
0.5
0.4
?
Langer
30
15
42
33
28.0
52.0
2.8
3.0
Langer
Feichter
Rasch
Barth
Koch
15
35.9
18
17
29
63
47.5
60.8
61.5
52.6
22.0
13.6
10.2
10.0
12.3
5.0
2.8
4.6
4.7
3.4
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
assumed to have the pH of the total cloud (Lin and Chameides
1991). Since cloud water typically has pH values less than 4,
H2O2 generally is considered to be the only oxidant because
rates of O3 oxidation of S(IV) decrease by 4 orders of magnitude between pH 5 and pH 4 (Savarino et al. 2000). The GCMs
that use this assumption also predict the lowest ⌬17O values. A
more refined analysis of cloud water S chemistry indicates that
this simplification is not valid for estimating which oxidant is
dominant in S(IV) oxidation (Lin and Chameides 1991; Chameides and Stelson 1992; Liang and Jacobson 1999). Lin and
Chameides pointed out that overall cloud pH is actually an
ensemble of two droplet populations, one acidic and one basic.
The acidic droplets form from H2SO4 nuclei generated by OH
oxidation of SO2, whereas the basic droplets arise from hydration of sea salt whose pH ranges from 6 to 8 in the marine
boundary layer (MBL). Modeling the pH of the droplets as
binned values, as opposed to bulk values, Lin and Chameides
(1991) calculated that 74% of in-cloud oxidation is due to O3
compared with 39% using the “bulk pH” approximation. Others
have reached similar conclusions and have shown that this
additional oxidation also increases the overall aqueous-phase
oxidation proportion of nss-SO4⫺2 (Liang and Jacobson 1999).
The importance of O3 oxidation of SO2 in wet sea-salt droplets
has also been inferred from field studies that found a larger
fraction of nss-SO4⫺2 in sea-salt particles than in smaller
particles associated with H2SO4 nuclei (O’Dowd, Lowe and
Smith 1999; Sievering et al. 1999).
Recalculations of the modeled ⌬17O values using box
models accounting for heterogeneous cloud chemistry and
the GCM estimates of OH oxidation are summarized in
Table 2. These values are in better agreement with the
observed desert SO4⫺2 values. The largest Atacama sulfate
⌬17O value (4 ‰) indicates that the Lin box model coupled
to either the Langer, Barth or Rasch GCM may provide the
best representation of S(IV) oxidation since these are the
only combinations that yield ⌬17O ⬎ 4‰. These model
results are also in good agreement with SO4⫺2 ⌬17O measurements from the Vostok ice core, which indicate nssSO4⫺2 ⌬17O fluctuations ranging from 1.3 to 4.8‰ over the
past 140000 yr (Alexander et al. 2002).
Using the modeled 5‰ value as the average nss-SO4⫺2
17
⌬ O value, we estimate that between 10 and 80% (average ⫽
34 ⫾ 24) of the SO4⫺2 associated with the nitrate ore deposits
in the Atacama Desert is derived from nss-SO4⫺2, with the
remainder from other sources such as sea salt or rock weathering. Direct comparison with data from the Vostok ice core,
where continental aerosols and sea salt are less than 10% of the
total SO4⫺2, could indicate between 30 and 100% of the
Atacama Desert SO4⫺2 is from atmospheric sources. It is
emphasized that the SO4⫺2 samples analyzed in this study are
from the nitrate ores and may not represent all of the SO4⫺2 in
the soil profiles within the ore fields and elsewhere in the region
(see section 3.3.1). The Mojave Desert SO4⫺2 samples appear
to have a slightly smaller average nss-SO4⫺2 component, but
the data are too limited to determine if they are systematically
different (Table 1).
The large spread in SO4⫺2 ⌬17O values (Fig. 3) indicates
that there are multiple SO4⫺2 sources and that some process has
limited homogenization of these sources in the depositional
environment. For SO4⫺2, with a dominant atmospheric ⌬17O
4031
value of ⬃5‰, variability among samples may be due to
mixing of SO4⫺2 with ⌬17O ⫽ 0 derived from sea salt, mineral
weathering, and volcanic sources that vary spatially. In addition, in contrast to NO3⫺atm, long-term variations in atmospheric SO4⫺2 ⌬17O values (Alexander et al. 2002) also may be
reflected in the variability of the ⌬17O found in the sulfate
minerals. Another possible explanation for SO42⫺ ⌬17O heterogeneity is selective leaching of the different aerosol types. Bao
et al. (2000b) proposed a selective leaching mechanism based
on aerosol size to explain SO4⫺2 ⌬17O variability with depth in
the Dry Valleys region of Antarctica. This mechanism may be
viable near the coast but we consider it to be less likely in areas
more than 10 km inland, especially in regions isolated by
mountains, where size segregation by gravitational settling has
already taken place (Wakshal and Nielsen 1982). An alternative
selective leaching model that is based on the chemical composition and reactivity of the individual aerosols is now suggested.
Submicron size NH4HSO4 aerosols, derived from gas phase
SO2 oxidation and H2O2 aqueous phase oxidation, are acidic
particles and would have correspondingly low ⌬17O values. In
the Atacama Desert boundary layer, where relative humidity is
often less than 20% (Cortes et al. 2000), these particles would
be crystalline and would be dry deposited. Submicron sea salt
aerosols are also present and are initially alkaline but will be
progressively acidified by reactions with HNO3 and by nssSO4⫺2 formation (dominated by O3 oxidation, with large
⌬17O). Nevertheless, even if the buffering capacity of the sea
salt is overwhelmed by acidification, the sea-salt-derived particles will still be neutral upon drying because of expulsion of
HCl gas. These two major SO4⫺2 aerosol types would have not
only different ⌬17O values, but also different deliquescence
points and reactivity. During the onset of a wetting event
(precipitation or fog), as relative humidity at the surface
reaches 40%, the NH4HSO4 will reach its deliquescence point
and form an acidic liquid droplet (Seinfeld and Pandis 1998). If
this droplet is on CaCO3 then dissolution of the CaCO3 will
occur with liberation of Ca2⫹ and CO2. Continued reaction
with CaCO3 eventually could raise the pH and promote volatilization of NH3, converting soluble NH4HSO4 to relatively
insoluble CaSO4:
NH4HSO4 → NH4⫹ ⫹ H⫹ ⫹ SO4⫺2
NH4⫹ → NH3(g) ⫹ H⫹
CaCO3 ⫹ 2H⫹ → Ca⫹2 ⫹ CO2(g) ⫹ H2O
Ca⫹2 ⫹ SO4⫺2 → CaSO4
net:
NH4HSO4 ⫹ CaCO3 → CaSO4 ⫹ CO2(g) ⫹ NH3(g) ⫹ H2O
(16)
In contrast, the sea salt particles are predominately NaCl, with
additional NaNO3 and NaSO4 from gas to particle conversion,
and sulfate minerals similar to marine evaporates. The deliquescence of these particles occurs at approximately 75% relative humidity (Seinfeld and Pandis 1998), but low Ca/SO4⫺2
ratios and lack of H⫹ limit the formation of insoluble CaSO4
and the more soluble sulfate salts are leached below the surface.
These contrasts in the surface behavior of different SO4⫺2
aerosols can account for several important observations: (i) the
paucity of carbonate in the soils of the Atacama Desert in
4032
G. Michalski, J. K. Böhlke, and M. Thiemens
comparison to other deserts (Berger and Cooke, 1997); (ii) the
abundance of CaSO4 at the surface (Rech et al. 2003); (iii) the
prevalence of sodium sulfate minerals in the nitrate ores (Searl
and Rankin 1993) and (iv) the increased ⌬17O values with
depth observed in polar deserts (Bao et al.2000b).
ⴚ2
4.3. Oceanic Versus Continental SO4 Aerosol
Deposition in the Atacama Desert
Given the isotopic evidence for large amounts of NO3⫺atm
and nss-SO4⫺ in Atacama Desert salts, along with meteorological evidence showing that air masses arriving in the Atacama
Desert can be either marine or continental in origin (Gallardo et
al.2002; Paluch et al. 1999) it is considered likely that there is
significant input of both marine and continental SO4⫺2 aerosols
to the desert surface. It is emphasized that marine aerosols are
not strictly sea salt particles, e.g., aerosols generated by sea
spray, but also include particles generated by atmospheric
chemistry in the MBL (nss-SO4⫺2, NO3⫺, organics). Sea salt
particles that are derived from bursting air bubbles at the sea
surface generate jet drops (diameter ⬎8 ␮m) and film droplets
(⬎3 ␮m). These are large particles and are rapidly removed by
gravitational settling, with the rates being a function of particle
size, altitude, and distance from the ocean. Although these large
particles make up the majority of sea salt on a mass basis, their
removal by dry and wet deposition on the Coastal Cordillera
would rapidly deplete their importance in the air mass reaching
the Atacama Desert. This phenomenon was clearly shown in
the ␦34S and Sr isotope data of Rech et al. (2003) and has been
used to interpret ␦34S gradients in SO4⫺2 in other coastal areas
(Wakshal and Nielsen 1982).
It is considered likely that the main component of marine
aerosols reaching the desert surface are particles roughly 1 ␮m
in diameter or smaller. Submicron particles are readily entrained into the free troposphere, where they can be transported
great distances, and can grow during cloud processing to micron size particles (the size fraction most efficient for HNO3
reactions) that are then removed by deposition. Submicron
SO4⫺2 aerosols obtained in the Atlantic at 23–27° S that were
considered a mixture of DMS oxidation products and submicron sea salt particles have ␦34S values of 7 to 15‰ (Patris et
al.2000). A similar study in the south Pacific gave ␦34S values
of ⬃ 15‰ for submicron SO4⫺2 aerosols (Calhoun et al.1991).
The primary sources of preindustrial continental sulfur, volcanic and biogenic emissions, have average ␦34S ⬃ 0‰ (Newman, Krouse, and Grinenko 1991). Therefore a model in which
roughly sub equal amounts of SO4⫺ are derived from continental aerosols (␦34S ⬃0‰) and submicron marine aerosols
(␦34S ⬃10 –15‰) is consistent with the S isotope data from our
study and that of Rech et al. (2003), and supports our hypothesis that most of the SO4⫺2 in the nitrate deposits in the
Atacama Desert is the result of atmospheric deposition.
5. SALT ACCUMULATION RATES IN THE
ATACAMA DESERT
Given isotopic evidence that the nitrate ores of the Atacama
Desert mainly represent long-term accumulation of atmospheric deposition, it is possible to estimate the amount of time
required for the deposits to form if estimates of the preindus-
trial atmospheric deposition rates (dry and wet) and the accumulated mass of the deposits are known.
5.1. Accumulation Rate of NO3ⴚatm by Dry Deposition
Nitrate accumulation by dry deposition of gas phase nitrogen
oxides and NO3⫺atm can be estimated based on known or
modeled deposition fluxes. Estimates of preindustrial N (and S)
concentrations based on the limited number of studies on
modern NOx or NO3⫺ aerosol concentrations in the Atacama
Desert, which are biased due to current anthroprogenic emissions (Saltzman et al.1986; Prado-Fiedler and Fuenzalida 1996;
Gallardo et al.2002), would have large uncertainties. With the
exception of polar regions and some midocean islands, continental air mass NO3⫺ concentrations are generally always
perturbed by human activity (Prospero and Savoie 1989). An
alternative approach is to use global N deposition models,
which are influenced by air mass origins, to derive estimates of
past NO3⫺atm deposition rates. The coastal range is dominated
by southerly winds driven by the Humboldt Current, however
the exact origin of the air masses reaching the interior desert is
less clear. Several studies suggest that air masses arriving in the
Atacama Desert are a mixture of oceanic and continental
(Paluch et al. 1999; Gallardo et al.2002). We will assume equal
proportions of oceanic and continental N deposition and models for each type of air mass will be utilized. The contribution
of N by direct deposition of NOx has been shown to be
insignificant in comparison to NO3⫺atm deposition and will not
be considered further (Duce et al.1991).
The global oceanic deposition model of Duce et al. (1991)
yielded deposition fluxes of ⬃20 mg of N m⫺2yr⫺1 from NO3⫺
for remote ocean islands and south Pacific surface waters, in
agreement with results given in Warneck (2000) for pristine
oceanic atmospheres. Approximately half this NO3⫺atm flux is
thought to be via dry deposition (⬃10 mg of N m⫺2yr⫺1).
Modeling of regional continental NO3⫺atm concentrations and
deposition rates is more complex because NO3⫺atm concentrations depend largely on the proximity of NOx sources and
deposition rates that vary with local meteorology and physiography (Dentener and Crutzen 1994; Adams et al. 1999; Holland
et al.1999; Levy et al.1999). In the preindustrial atmosphere,
the main NOx sources were soil emissions from denitrification,
nitrification, natural biomass burning, lightning, and stratospheric mixing of NOx generated by N2O photolysis (Holland
et al.1999). The respective magnitudes of these sources would
have been influenced by the presence of the Amazon basin
(biomass burning, soil exhalation) and the Andean range
(stratospheric mixing, lightning), which may have given rise to
additional NO3⫺atm influxes to the Atacama Desert. A detailed
estimate of the relative importance of these regions for local
NO3⫺atm fluxes is difficult because of uncertainties in the
amount and type of biomass burning and overall uncertainties
in lightning and stratospheric fluxes (Holland et al. 1999). In
addition, the effective removal by wet and dry deposition in the
Andean range of NO3⫺atm produced in the Amazon basin
should limit the influence of this region on NO3⫺atm reaching
the Atacama. These factors were considered in the global
nitrogen deposition model of Holland et al. (1999) in which
estimated global preindustrial nitrogen fluxes in continental
regions were based on land and plant types, local emission
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
inventories, and global circulation and precipitation. Nitrate
fluxes to desert regions in the southern hemisphere were estimated to be between 21 and 84 mg N m⫺2 yr⫺1 with an average
of 44 mg N m⫺2 yr⫺1, with approximately half of this flux
attributed to dry deposition (22 mg N m⫺2 yr⫺1). If marine and
continental air masses are about equally represented on average
in the region, then these deposition models would imply a total
dry deposition NO3⫺ flux to the Atacama Desert surface of 16
mg N m2 yr⫺1.
5.2. Accumulation Rate of NO3ⴚatm by Wet Deposition
Because measurable precipitation in the Atacama Desert can
occur as rarely as once a century (Ericksen 1983), the contribution of NO3⫺atm from rainout or washout is considered
negligible. However, the upwelling of nutrient rich, cold bottom water off the Chilean coast intensifies the aridity in the
Atacama by condensing water vapor from oceanic air and
generates heavy fogs called “camanchaca.” Camanchaca occurs
up to 189 d annually in some areas (Cereceda and Schemenauer
1991), with central Chile averaging 45 d, and the fog has been
repeatedly suggested as a possible source of the NO3⫺ in the
nitrate deposits (Searl and Rankin 1993; Berger and Cooke
1997; Rech et al.2003). In latitudes south of 30° S, the occurrence of fog decreases because of atmospheric subsidence
created by the South Pacific anticyclone. This, combined with
the Coastal Cordillera barrier, limits fog reaching the region of
the nitrate deposits to only ⬃15 d annually (Cereceda and
Schemenauer 1991). To evaluate this potential source of dissolved ions in the Atacama Desert requires an estimation of fog
water deposition rates and the concentrations of soluble ions in
the fog water.
The virtual absence of vegetation in the region simplifies
deposition velocities to gravitational settling since resistance
from leaves and shrubs is zero. For fog droplets having a
typical diameter of ⬃ 10 ␮m, a liquid water content of 1.5 g
m⫺3 and a deposition velocity of 3 cm s⫺1, the amount of fog
water deposited to the desert surface would be ⬃0.2 L day⫺1
m⫺2 (Seinfeld and Pandis 1998). This rate will be slightly
higher for hillsides, but for our purposes, a flat terrain approximation is satisfactory. A single fog-water sample collected east
of Antofagasta, Chile contained Cl⫺, NO3⫺, and SO4⫺2 concentrations of 46, 19, and 32 mg/kg, respectively (Ericksen
1981). However, in a separate study at El Tofo (inner coastal
range) these same anion concentrations in fog were only 4.5,
0.75 and 5.9 mg/kg (Schemenauer and Cereceda 1992). Fog
water from both sites contained detectable amounts of sea salt,
but the authors also noted the possible influence of anthroprogenic emissions on NO3⫺ and SO4⫺2 concentrations, as well as
the possible contribution of local dust that is highly enriched in
both NO3⫺ and SO4⫺2. Vong et al. (1997) analyzed northern
hemisphere marine cloud water and found Cl⫺, NO3⫺, and
SO4⫺2 concentrations of 15, 0.9, and 4.6 mg/kg, which they
considered to represent background marine cloud concentrations. Much of the fog invading the inland Chilean valleys is
orographic fog, originating from stratocumulus clouds in the
Pacific Ocean intercepting the coastal range (Cereceda et al.
2002). This observation, combined with the lack of data from
Chilean fog samples unaffected by local or distant anthroprogenic sources, prompts us to use the marine cloud-water con-
4033
centrations of Vong et al. (1997). These data result in a fog
NO3⫺ deposition flux of approximately 0.5 mg N m2 yr⫺1,
which is much smaller than the NO3⫺atm dry deposition flux.
5.3. Accumulation Time Estimates for NO3
The Atacama Desert nitrate ore deposits are located primarily on the eastern side of the Coastal range in an area that is
approximately 700 km long and 30 km wide, which would
represent the minimum deposition area (21000 km2) (Ericksen
1983). A larger area encompassing the entire watershed could
be considered as the deposition area if it is assumed that the
NO3⫺ was redistributed and concentrated locally by hydrologic
or eolian processes. The total estimated wet (fog) and dry
deposition flux of 16.5 mg N m2yr⫺1 (as NO3⫺) spread over
the estimated area of the nitrate ore deposits (21000 km2) could
account for ⬃350 metric tons of NO3⫺-N annually. The
amount of high-grade nitrate ore in the Atacama Desert is
estimated to have been around 33 million metric tons (as N)
and the more widespread low-grade deposits are perhaps twice
this amount (Ericksen 1983), for a total of approximately 100
Mt as N. The ⌬17O data indicate that approximately 75% of the
total (75 million tons of NO3⫺-N) is derived directly from
NO3⫺atm deposition, which would require a minimum of
214,000 yr to accumulate in the area of the major deposits. The
required time would be shorter if the NO3⫺ was accumulated
over a larger area and then concentrated in the deposits via
postdeposition redistribution, or it could be longer if the estimates of NO3⫺ reserves are biased to exclude large amounts of
material with low (non-economic) concentrations.
The hyper-arid conditions of the Atacama Desert are estimated to have existed for at least 6 million years and perhaps
as long as 15 million years (Alpers and Brimhall 1988; Hartley
and Chong 2002), which is more than sufficient time to accommodate our deposition estimates. It is possible that the 214,000
yrs. underestimates by a factor of two, or perhaps even by an
order of magnitude, the actual time required to accumulate the
NO3⫺ because of uncertainties in the deposition rates, the air
masses trajectories and estimates of the NO3⫺ reservoir in so
large an area. In addition, this timeframe is based on the
assumption that all deposited NO3⫺ is retained in the soil and
that no net loss by wind erosion or denitrification has occurred.
␦15N values indicate that denitrification is not a major sink for
NO3⫺ in the Atacama Desert (Böhlke et al.1997), which is
reasonable given the lack of water, yet wind erosion in deserts
is considered significant (Reheis and Kihl 1995). The importance of wind erosion for NO3⫺atm accumulation depends on
the rate at which NO3⫺ is leached and transported downward
below the exposed surface. The fogs and rare precipitation
events, which only contribute a small fraction of the total
NO3⫺atm deposition rate, may play an important role in moving
NO3⫺ below the surface. The arid conditions also are punctuated by brief periods of heavy rainfall (Vargas et al. 2000) and
perhaps longer periods (103 to 104 years or longer) of overall
decreased aridity. These conditions may have helped create
vertical salt zonation based on solubility differences, and they
may have periodically diminished the NO3⫺ salt deposits by
leaching them into ground water or to deeper horizons in the
unsaturated zone. If our estimates are correct, then it is possible
that the modern (preexploitation) nitrate ore fields represented
4034
G. Michalski, J. K. Böhlke, and M. Thiemens
only a small fraction of the total accumulation of atmospheric
NO3⫺ since hyperarid conditions were established in the region
and that large reservoirs of NO3⫺ either exist in the subsurface
or have been removed.
5.4. Accumulation Rates of Other Salts in the
Atacama Desert
The isotopic evidence and NO3⫺atm deposition calculations
indicate that the accumulation of atmospheric particles is the
major source of the NO3⫺ and SO4⫺2 in the Atacama Desert
nitrate ore fields. An additional check on this conclusion is to
estimate what the expected anion ratios should be for atmospheric deposition and compare those ratios with the observed
ratios in the deposits. Submicron aerosols in clean marine air
contain approximately equal molar amounts of Cl⫺, sea salt
SO4⫺2, and nss-SO4⫺2, but only about one tenth as much
NO3⫺atm (Sievering et al. 1999). Preanthroprogenic continental
aerosol anion ratios are more difficult to assess because of the
anthroprogenic influences in modern studies. One approach is
to adopt anion ratios from Andean ice core data. These cores
span the past 20000 yr and represent the anion composition of
high-elevation, east-to-west moving air masses. Anion data
from the Sajama ice cap in the Bolivian Andes east of the
Atacama Desert exhibit some variability on 1000 yr timescales,
but average molar ratios of NO3⫺/Cl⫺ and NO3⫺/SO4⫺2 are
roughly 2:1 and 1:2 respectively (Thompson et al. 1998;
Thompson et al. 2000). As a first approximation, assuming half
continental and half oceanic sources for the air masses (see
above), one might expect bulk ratios of NO3⫺/Cl⫺ and NO3⫺/
SO4⫺2 to be ⬃1.0 and 0.3, respectively, in the Atacama Desert.
Data compiled by Grossling and Ericksen (1971) and Ericksen (1981) indicate NO3⫺/Cl⫺ and NO3⫺/SO4⫺2 molar ratios
in the bulk nitrate ores are around 1.27 and 1.03, respectively
(Table 3). The apparent relative enrichment of NO3⫺ over Cl⫺
and SO4⫺2 in the ores may simply be an artifact since the salt
ratios compiled for the nitrate ore layers may not represent the
total salt ratios in the region. For example, failure to include
some of the non-soluble sulfate phases such as CaSO4 that
occur above the nitrate ores could explain the difference in the
Table 3. Average composition of soluble salts in nitrate ores of the
Atacama Desert (data from Grossling and Ericksen, 1971) derived from
815 analyses representing monthly averages from 1932 to 1967 for ores
treated by the two largest nitrate production plants (Oficina Pedro de
Valdivia and Oflcina Maria Elena). Relatively insoluble phases containing Ca⫹2, SO4⫺2, B5O9⫺3, and IO3⫺ may be underestimated.
Soluble
component
Fraction
in ore
(wt. %)
Mole ratio
with respect to
Cl⫺
Cl⫺
SO4⫺2
NO3⫺
B5O9⫺3
IO3·
ClO4⫺
Na⫹
Ca⫹2
Mg⫹2
K⫹
Sum
4.58
10.00
6.29
0.50
0.060
0.028
6.90
1.81
0.46
0.73
31.36
1
0.806
0.786
0.020
0.0027
0.0022
2.33
0.35
0.15
0.14
NO3⫺/SO4⫺2 ratios. Another uncertainty is the degree to which
wind erosion has fractionated the various salts. Because Cl⫺
and NO3⫺ are more readily leached below the surface, surficial
gypsum and anhydrite may be more susceptible to erosion, with
the result that SO4⫺2 may be selectively transported out of the
region by wind. Therefore, the bulk anion composition of the
ores is considered reasonably consistent with an atmospheric
source for much of the NO3⫺, SO4⫺2, and Cl⫺, as suggested by
Böhlke et al. (1997). Acknowledging the atmospheric source of
NO3⫺ while simultaneously discounting the importance of
nss-SO4⫺2 and Cl⫺ deposition (Rech et al. 2003) would appear
to be unsupported by these data.
Substantial amounts of iodate and perchlorate salts also are
present in the Atacama Desert nitrate ores. Grossling and
Ericksen (1971) estimate that the average bulk nitrate ores have
molar ratios of IO3⫺/NO3⫺ ⫽ 0.0034; ClO4⫺/NO3⫺ ⫽ 0.0028
(Table 3). Iodate is the only aerosol iodine species that is not
volatile and it is formed through gas phase oxidation of IO by
O3 in a process that resembles the NOx cycle (Chatfield and
Crutzen 1990). There is little anthroprogenic IO3⫺ in comparison to natural IO3⫺ in modern aerosols, which have IO3⫺
concentrations ranging from 1 to 7 ng m⫺3 (Wimschneider and
Heumann 1995; Baker et al. 2001). Combining those concentrations with preindustrial estimates of ⬃ 20 ng m⫺3 of
NO3⫺atm (Duce et al. 1991; Warneck 2000) would give IO3⫺/
NO3⫺ molar ratios of 0.002 to 0.013, similar to the bulk ratios
in the Atacama Desert nitrate ores, consistent with the IO3⫺ in
the ores being mainly the result of photochemistry and subsequent atmospheric deposition. Because of its interaction with
O3, the O isotopic composition of IO3⫺ might indicate its
origin; however, the absence of significant ⌬17O would not
necessarily exclude photochemistry because exchange between
IO3⫺ and water is rapid (Gramsjäger and Murmann 1983).
Perchlorate is perhaps even more interesting with respect to
the origins of the salts. No measurements of ClO4⫺ concentrations in aerosol have been made to our knowledge. However,
several authors have suggested the possibility of stratospheric
production of aerosol ClO4⫺ through the reaction of perchloric
acid or ClO on stratospheric SO4⫺2 aerosols (Jaegle et al.1996;
Murphy and Thomson 2000). ClO chemistry is remarkably
similar to NOx chemistry, with rapid photolysis and O3 oxidation sequences, which prompted Lyons to propose the existence
of substantial ⌬17O values in stratospheric ClO (Lyons 2001).
There is no known source of tropospheric ClO4⫺ or other
geologic mechanism known to produce substantial amounts of
ClO4⫺. Therefore, if the O isotopic composition of ClO4⫺ salts
could be shown to have non-zero ⌬17O values, then it might be
concluded either that some of the Atacama Desert salts are
from stratospheric aerosols (Ericksen 1983) or that there is an
unknown source of tropospheric or terrestrial ClO4⫺. Investigations on the isotopic composition of both these compounds
should be carried out.
6. CONCLUSIONS
The O isotopic composition (␦18O, ␦17O, and ⌬17O) of
NO3⫺ and SO4⫺2 in the nitrate mineral deposits of the
Atacama Desert have confirmed that the main source of
these anions is the long term accumulation of NO3⫺and
SO4⫺2 produced by photochemical reactions in the atmo-
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
sphere. The deposits represent accumulation of submicron
particles transported in the free troposphere and removed
primarily by dry deposition. Variation in NO3⫺ ⌬17O values
is most likely the result of mixing of NO3⫺ derived from
microbial nitrification with the atmospheric photochemical
source. Comparative data from the Mojave Desert indicate
that the photochemical NO3⫺ isotope signature is well preserved in the driest surficial environments that are almost
lifeless, whereas the microbial NO3⫺ isotope signature becomes dominant rapidly with increasing moisture, biologic
activity, and N cycling. The SO4⫺2 isotope data indicate a
large fraction of the SO4⫺2 is derived from atmospheric
nss-SO4⫺2 that is produced by heterogeneous and homogeneous oxidation of SO2. However, a more complete study of
SO4⫺2 O isotopes based on location and depth should be
undertaken to assess the importance of aerosol deliquescence and other surface reaction properties on SO4⫺2 ⌬17O
values.
These findings are important for a variety of fields in
biogeochemistry. In the Atacama Desert, ⌬17O values represent a preserved isotopic signature of atmospheric processes that occurred over hundreds of thousands to millions
of years. If future isotopic measurements could be coupled
with detailed dating studies, it is possible that new insights
would be obtained about atmospheric processes dating back
millions of years, where few proxies are currently available.
In paleoclimate studies using ice cores, where in situ nitrification is limited and annual deposition is isolated in ice
layers, NO3⫺ ⌬17O variations might indicate shifts in atmospheric oxidative pathways. This principle also may be
applied to extraterrestrial planetary studies where, for example, investigations of atmosphere/regolith interactions utilizing mass independent isotopic compositions have raised new
questions about the evolution of the Martian atmosphere
(Farquhar et al. 2000; Farquhar and Thiemens 2000). The
existence of NO3⫺ in the Martian regolith has been proposed
theoretically (Yung and DeMore 1999) and confirmed by
analyses (Grady et al.1995) and isotopic analyses of future
returned samples may provide information on the ancient
Martian atmosphere unattainable by other means.
Measurements of the complete O isotopic composition of
NO3⫺ are useful not only for tracing atmospheric/photochemical processes but also for probing the influence of
biologic processes on NO3⫺ formation. In natural systems
on Earth, NO3⫺ formation is mainly the result of photochemistry (⌬17O ⬎ 0) or biologic nitrification (⌬17O ⫽ 0).
The Atacama Desert is commonly studied as a model for the
Martian surface, and it is possible that ⌬17O measurements
could be used to identify products of nitrification by biologic
processes in extraterrestrial samples as well, making this a
potentially important planetary biomarker. On Earth, NO3⫺
⌬17O values can provide a measure of the contribution of
NO3⫺atm deposition in aquatic systems where a variety of
natural and anthropogenic NO3⫺ sources are present. For
example, this would be useful in northern hemisphere countries concerned with the effects of N deposition associated
with fossil fuel burning. This study has highlighted the
importance of examining all three stable isotopes of oxygen
in studies concerned with interactions between atmospheric
and terrestrial/biosphere processes.
4035
Acknowledgments— This project was funded in part by the National
Science Foundation and the USGS National Research Program in
Water Resources. Assistance in the USGS laboratory was provided by
Tyler Coplen, Janet Hannon, Stan Mroczkowski, and Haiping Qi. We
would like to thank Martin Miller, Kinga Revesz and an anonymous
reviewer for their helpful comments.
Associate editor: D. Cole
REFERENCES
Adams P. J., Seinfeld J. H. and Koch D. M. (1999) Global concentrations of tropospheric sulfate, nitrate, and ammonium aerosol simulated in a general circulation model. J. Geophys. Res. 104, 13791–
13823.
Alexander B., Savarino J., Barkov N. I., Delmas R. J. and Thiemens
M. H. (2002) Climate driven changes in the oxidation pathways of
atmospheric sulfur. Geophys. Res. Lett. 29, 1681–1685.
Alpers C. N. and Brimhall G. H. (1988) Middle Miocene climaticchange in the Atacama Desert, northern Chile— evidence from
supergene mineralization at La-Escondida. Geol. Soc. Am. Bull.
100, 1640 –1656.
Aravena R., Suzuki O., Pena H., Pollastri A., Fuenzalida H. and Grilli
A. (1999) Isotopic composition and origin of the precipitation in
northern Chile. Appl. Geochem. 14, 411– 422.
Baker A. R., Tunnicliffe C. and Jickells T. D. (2001) Iodine speciation
and deposition fluxes from the marine atmosphere. J. Geophys. Res.
106, 28743–28749.
Bao H., Michalski G. M. and Thiemens M. H. (2001) Sulfate oxygen-17 anomalies in desert varnishes. Geochim. Cosmochim. Acta
65, 2029 –2036.
Bao H. and Thiemens M. H. (2000b) Generation of O2 from BaSO4
Using a CO2-Laser Fluorination System for Simultaneous Analysis
of ␦18O and ␦17O. Anal. Chem. 72, 4029 – 4032.
Bao H., Campbell D. A., Bockheim J. G. and Thiemens M. H. (2000a)
Origins of sulphate in Antarctic dry-valley soils as deduced from
anomalous 17O compositions. Nature 407, 499 –502.
Bao H., Thiemens M. H., Farquhar J., Campbell D. A., Lee C. C. W.,
Heine K. and Loope D. B. (2000) Anomalous 17O compositions in
massive sulfate deposits on the Earth. Nature 406, 176 –178.
Barth M. C., Rasch P. J., Kiehl J. T., Benkovitz C. M., and Schwartz
S. E. (2000) Sulfur chemistry in the National Center for Atmospheric Research Community Climate Model: Description, evaluation, features, and sensitivity to aqueous chemistry. J. Geophys.
Res. 105, 1387–1415.
Berger I. A. and Cooke R. U. (1997) The origin and distribution of salts
on alluvial fans in the Atacama Desert, northern Chile. Earth Surf.
Proc. and Landforms 22, 581– 600.
Böhlke J. K., Ericksen G. E. and Revesz K. (1997) Stable isotope
evidence for an atmospheric origin of desert nitrate deposits in
northern Chile and southern California, USA. Chem. Geol. 136,
135–152.
Böhlke J. K., Mroczkowski S. J., and Coplen T. B. (2003) Oxygen
isotopes in nitrate: new reference materials for 18O:17O:16O measurements and observations on nitrate-water equilibration. Rapid
Comm. Mass. Spec. 17, 1835–1846.
Calhoun J. A., Bates T. S. and Charlson R. J. (1991) Sulfur isotope
measurements of submicrometer sulfate aerosol particles over the
Pacific Ocean. Geophys. Res. Lett. 18, 1877–1880.
Casciotti K. L., Sigman D. M., Hastings M. G., Böhlke J. K., and
Hilkert A. (2002) Measurement of the oxygen isotopic composition
of nitrate in seawater and freshwater using the denitrifier method.
Anal. Chem. 74, 4905– 4912.
Cereceda P., Osses P., Larrain H., Farias M., Lagos M., Pinto R. and
Schemenauer R. S. (2002) Advective, orographic and radiation fog
in the Tarapaca region, Chile. Atmos. Res. 64, 261–271.
Cereceda P. and Schemenauer R. S. (1991) The occurrence of fog in
Chile. J. Appl. Meteor. 30, 1097–1105.
Chameides W. L. and Stelson A. W. (1992) Aqueous-phase chemical
processes in deliquescent sea-salt aerosols: A mechanism that couples the atmospheric cycles of sulfur and sea salt. J. Geophys. Res.
97, 20,565–20,580.
4036
G. Michalski, J. K. Böhlke, and M. Thiemens
Chatfield R. B. and Crutzen P. J. (1990) Are there interactions of iodine
and sulfur species in marine air photochemistry? J. Geophys. Res.
95, 22319 –22341.
Chin M., Rood R. B., Lin S. J., Muller J. F., and Thompson A. M.
(2000) Atmospheric sulfur cycle simulated in the global model
GOCART: Model description and global properties. J. Geophys.
Res. 105, 24671–24687.
Chong G. (1994) The Nitrate Deposits of Chile. In Tectonics of the
Southern Central Andes: Structure and Evolution of an Active
Continental Margin (ed. K. J. Reutter, E. Scheuber and P. J.
Wigger), pp. 303–316. Springer-Verlag, Berlin.
Claridge G. G. C. and Campbell I. B. (1968) Origin of nitrate deposits.
Nature 217, 428 – 430.
Cortes A., Miranda E., Rosenmann M. and Rau J. R. (2000) Thermal
biology of the fossorial rodent Ctenomys fulvus from the Atacama
Desert, northern Chile. J. Therm. Bio. 25, 425– 430.
Darwin C. (1871) The Journal of a Voyage in H. M. S. Beagle. D.
Appleton and Co., London
Densmore J. N. and Böhlke J. K. (2000) Use of nitrogen isotopes to
determine sources of nitrate contamination in two desert basins in
California. In Interdisciplinary Perspectives on Drinking Water
Risk Assessment and Management, International Association of
Hydrological Sciences (ed. Reichard, E. G., Hauchman, F. S., and
Sancha, A. M.). 20, 63–73.
Dentener F. J. and Crutzen P. J. (1993) Reaction of nitrogen pentoxide
on tropospheric aerosols: Impact on the global distributions of NOx,
ozone, and hydroxyl. J. Geophys. Res. 98, 7149 –7163.
Dentener F. J. and Crutzen P. J. (1994) A 3-Dimensional model of the
global ammonia cycle. J. Atmos. Chem. 19, 331–369.
Derwent R. G., Jenkin M. E., Johnson C. E., and Stevenson D. S.
(2003) The global distribution of secondary particulate matter in a
3-D Lagrangian chemistry transport model. J. Atmos. Chem. 44,
57–95.
Dubey M. K., Mohrschladt R., Donahue N. M. and Anderson J. G.
(1997) Isotope-specific kinetics of hydroxyl radical (OH) with
water (H2O): Testing models of reactivity and atmospheric fractionation. J. Phys. Chem. A. 101, 1494 –1500.
Duce R. A., Liss P. S., Merrill J. T., Atlas E. L., Buat-Menard P., Hicks
B. B., Miller J. M., Prospero J. M. and Arimoto R. (1991) The
atmospheric input of trace species to the world ocean. Global
Biogeochem. Cycles 5, 193–259.
Ericksen G. E. (1979) Origin of the Nitrate Deposits of Northern Chile,
Segundo Congreso Geologico Chileno. Conference proceeding.
Arica, Chile. c181-c205.
Ericksen G. E (1981) Geology and Origin of the Chilean Nitrate
Deposits. Professional paper 1188, 37. USGS.
Ericksen G. E. (1983) The Chilean nitrate deposits. Am. Sci. 71,
366 –374.
Ericksen G. E., Hosterman J. W., and Stamand P. (1988) Chemistry,
mineralogy, and origin of the Clay-Hill nitrate deposits, Amargosa
river valley, Death Valley region, California, USA. Chem. Geol. 67,
85–102.
Farquhar J., Bao H., and Thiemens M. (2000) Atmospheric influence of
Earth’s earliest sulfur cycle. Science 289, 756 –758.
Farquhar J., Savarino J., Jackson T. L. and Thiemens M. H. (2000)
Evidence of atmospheric sulphur in the Martian regolith from
sulphur isotopes in meteorites. Nature 404, 50 –52.
Farquhar J. and Thiemens M. H. (2000) Oxygen cycle of the Martian
atmosphere-regolith system: ⌬17O of secondary phases in Nakhla
and Lafayette. J. Geophys. Res. 105, 11991–11997.
Feichter J., Kjellstrom E., Rodhe H., Dentener F., Lelieveld J., and
Roelofs G. J. (1996) Simulation of the tropospheric sulfur cycle in
a global climate model. Atm. Envir. 30, 1693–1707.
Gallardo L., Olivares G., Langner J. and Aarhus B. (2002) Coastal lows
and sulfur air pollution in central Chile. Atmos. Environ. 36, 3829 –
3841.
Gao Y. Q. and Marcus R. A. (2001) Strange and unconventional
isotope effects in ozone formation. Science 293, 259 –263.
Gong S. L., Barrie L. A., Prospero J. M., Savoie D. L., Ayers G. P.,
Blanchet J. P. and Spacek L. (1997) Modeling sea-salt aerosols in
the atmosphere, 2. Atmospheric concentrations and fluxes J. Geophys. Res. 102, 3819 –3830.
Grady M. M., Wright I. P. and Pillinger C. T. (1995) A search for
nitrates in Martian meteorites. J. Geophys. Res. 100, 5449 –5455.
Gramsjäger H. and Murmann R. K. (1983) Oxygen-18 Exchange
Studies of Aqua- and Oxo-ions. In Advances in Inorganic and
Bioinorganic Mechanisms Vol. 2 (ed. A.G. Sykes), pp. 317–380.
Academic Press, London.
Hartley A. J. and Chong G. (2002) Late. Pliocene age for the Atacama
Desert: Implications for the desertification of western South America. Geology 30, 43– 46.
Hathorn B. C. and Marcus R. A. (1999) An intramolecular theory of the
mass-independent isotope effect for ozone. I. J. Chem. Phys. 111,
4087– 4100.
Holland E. A., Dentener F. J., Braswell B. H. and Sulzman J. M. (1999)
Contemporary and pre-industrial global reactive nitrogen budgets.
Biogeochem. 46, 7– 43.
Hollocher T. C. (1984) Source of the oxygen-atoms of nitrate in the
oxidation of nitrate by nitrobacter-agilis. Arch. Biochem. Biophys.
233, 721–727.
Jaegle L., Yung Y. L., Toon G. C., Sen B. and Blavier J. F. (1996)
Balloon observations of organic and inorganic chlorine in the
stratosphere: The role of HClO4 production on sulfate aerosols.
Geophys. Res. Lett. 23, 1749 –1752.
Johnston J. C. and Thiemens M. H. (1997) The isotopic composition of
tropospheric ozone in three environments. J. Geophys. Res. 102,
25395–25404.
Kendall C. (1998) Tracing Nitrogen Sources and Cycling in Catchments. In Isotope Tracers in Catchment Hydrology (ed. C.
Kendall and J. J. McDonnell), pp. 519 –576. Elsevier Science,
Amsterdam.
Koch D., Jacob D., Tegen I., Rind D., and Chin M. (1999) Tropospheric sulfur simulation and sulfate direct radiative forcing in the
Goddard Institute for Space Studies general circulation model. J.
Geophys. Res. 104, 23799 –23822.
Kornexl B. E., Gehre M., Hofling R. and Werner R. A. (1999) On-line
␦18O measurement of organic and inorganic substances. Rap. Commun. Mass Spectr. 13, 1685–1693.
Krankowsky D., Bartecki F., Klees G. G., Mauersberger K., Schellenbach K. and Stehr J. (1995) Measurement of heavy isotope enrichment in tropospheric ozone. Geophys. Res. Lett. 22, 1713–1716.
Lammerzahl P., Rockmann T., Brenninkmeijer C. A. M., Krankowsky
D., and Mauersberger K. (2002) Oxygen isotope composition of
stratospheric carbon dioxide. Geophys. Res. Lett. 29, Art. No. 1582.
Langner J. and Rodhe H. (1991) A global three-dimensional model of
the tropospheric sulfur cycle. J. Atmos. Chem. 13, 225–263.
Leatham S, Huckins L. N, Jacobson R. L (1983) Nitrate in Nevada
playas. Desert Research Institution, University of Nevada, Reno.
Publication. 41086.
Lee C. C. W. and Thiemens M. H. (2001) The ␦17O and ␦18O
measurements of atmospheric sulfate from a coastal and high
alpine region: A mass- independent isotopic anomaly. J. Geophys. Res. 106, 17359 –17373.
Lee C. C. W. (2001) Mass independent oxygen isotopic composition of
atmospheric sulfate: Origin and implications for the present and
past atmosphere of Earth and Mars. Geophys. Res. Lett. 28, 1783.
Lee C. C. W (2000) Multiple stable oxygen isotopic studies of atmospheric sulfate: a new quantitative way to understand sulfate formation processes in the atmosphere. Ph.D. Dissertation. University
of California, San Diego.
Levy H. II, Moxim W. J., Klonecki A. A. and Kasibhatla P. S. (1999)
Simulated tropospheric NOx: its evaluation, global distribution and
individual source contributions. J. Geophys. Res. 104, 26279 –
26306.
Liang J. Y. and Jacobson M. Z. (1999) A study of sulfur dioxide
oxidation pathways over a range of liquid water contents, pHs, and
temperatures. Adv. Air Poll. 6, 989 –996.
Lin X. and Chameides W. L. (1991) Model studies of the impact of
chemical inhomogeneity on sulfur dioxide oxidation in warm stratiform clouds. J. Atmos. Chem. 13, 109 –129.
Logan J. A., Prather M. J., Wofsy S. C. and McElroy M. B. (1981)
Tropospheric chemistry—a global perspective. J. Geophys. Res. 86,
7210 –7254.
Deposition of nitrate in the Atacama desert: Evidence from ⌬17O
Lyons J. R. (2001) Transfer of mass-independent fractionation in ozone
to other oxygen-containing radicals in the atmosphere. Geophys.
Res. Lett. 28, 3231–3234.
Matsuhisa Y., Goldsmith J. R., and Clayton R. N. (1978) Mechanisms
of hydrothermal crystallization of quartz at 250°C and 15 kbar.
Geochim. Cosmochim. Acta 42, 173–182.
Mauersberger K., Erbacher B., Krankowsky D., Gunther J., and Nickel
R. (1999) Ozone isotope enrichment: isotopomer-specific rate coefficients. Science 283, 370 –372.
Meijer H. A. J. and Li W. J. (1998) The use of electrolysis for accurate
␦17O and ␦18O isotope measurements in water. Iso. Env. Health
Stud. 34, 349 –369.
Michalski G., Savarino J., Böhlke J. K. and Thiemens M. (2002)
Determination of the total oxygen isotopic composition of nitrate
and the calibration of a ⌬17O nitrate reference material. Anal.
Chem. 74, 4989 – 4993.
Michalski G., Scott Z., Kabiling M. and Thiemens M. (2003) First
measurements and modeling of ⌬17O in atmospheric nitrate. Geophys. Res. Lett. 30[16], ASC14 –1. 2003.
Miller M. F. (2002) Isotopic fractionation and the quantification of
O-17 anomalies in the oxygen three-isotope system: an appraisal
and geochemical significance. Geochim. Cosmochim. Acta 66,
1881–1889.
Morton J., Barnes J., Schueler B. and Mauersberger K. (1990) Laboratory studies of heavy ozone. J. Geophys. Res. 95, 901–907.
Mueller G. (1968) Genetic histories of nitrate deposits from Antarctica
and Chile. Nature 219, 1131–1134.
Gramsjäger H. and Murmann R. K. (1983) Oxygen-18 Exchange
Studies of Aqua- and Oxo-ions. In Advances in Inorganic and
Bioinorganic Mechanisms Vol. 2 (ed. A.G. Sykes), pp. 317–380.
Academic Press, London.
Murphy D. M. and Thomson D. S. (2000) Halogen ions and NO⫹ in the
mass spectra of aerosols in the upper troposphere and lower stratosphere. Geophys. Res. Lett. 27, 3217–3220.
Newman L., Krouse H. R. and Grinenko V. A. (1991) Sulfur Isotope
Variations in the Atmosphere. In Stable Isotopes: Natural and
Anthropogenic Sulphur in the Environment (ed. H. R. Krouse and
V. A. Grinenko), pp. 440. Wiley, New York.
O’Dowd C. D., Lowe J. A. and Smith M. H. (1999) Observations and
modeling of aerosol growth in marine stratocumulus-case study.
Atmos. Environ. 33, 3053–3062.
O’Dowd C. D., Lowe J. A., Clegg N., Smith M. H., and Clegg S. L.
(2000) Modeling heterogeneous sulphate production in maritime
stratiform clouds. J. Geophys. Res. 105, 7143–7160.
Patris N., Mihalopoulos N., Baboukas E. D. and Jouzel J. (2000)
Isotopic composition of sulfur in size-resolved marine aerosols
above the Atlantic Ocean. J. Geophys. Res. 105, 14449 –14457.
Peierls B. L. and Paerl H. W. (1997) Bioavailability of atmospheric
organic nitrogen deposition to coastal phytoplankton. Limnol.
Oceanogr. 42, 1819 –1823.
Prado-Fiedler R. and Fuenzalida H. A. (1996) Wet and dry deposition
of nitrogen compounds in the southeast Pacific coast: Montemar,
central Chile. J. Geophys. Res. 101, 22845–22853.
Prospero J. M. and Savoie D. L. (1989) Effect of continental sources on
nitrate concentrations over the Pacific Ocean. Nature 339, 687–
689.
Pueyo J. J., Chong G. and Vega M. (1998) Mineralogy and parental
brine evolution in the Pedro de Valdivia nitrate deposit, Antofagasta, Chile. Rev. Geolog. Chile 25, 3–15.
Rasch P. J., Barth M. C., Kiehl J. T., Schwartz S. E., and Benkovitz
C. M. (2000) A description of the global sulfur cycle and its
controlling processes in the National Center for Atmospheric Research Community Climate Model, Version 3. J. Geophys. Res.
105, 1367–1385.
Rech J. A., Quade J. and Hart W. S. (2003) Isotopic evidence for the
source of Ca and S in soil gypsum, anhydrite and calcite in the
Atacama Desert, Chile. Geochim. Cosmochim. Acta 67, 575–
586.
Reheis M. C. and Kihl R. (1995) Dust deposition in southern Nevada
and California, 1984 –1989 —relations to climate, source area, and
source lithology. J. Geophys. Res. 100, 8893– 8918.
4037
Revesz K. and Böhlke J. K. (2002) Comparison of ␦18O measurements
in nitrate by different combustion techniques. Anal. Chem. 74,
5410 –5413.
Revesz K., Böhlke J. K. and Yoshinari T. (1997) Determination of
␦18O and ␦15N in nitrate. Anal. Chem. 69, 4375– 4380.
Savarino J., Lee C. C. W. and Thiemens M. H. (2000) Laboratory
oxygen isotopic study of sulfur (IV) oxidation: Origin of the massindependent oxygen isotopic anomaly in atmospheric sulfates and
sulfate mineral deposits on Earth. J. Geophys. Res. 105, 29079 –
29088.
Savarino J. and Thiemens M. H. (1999) Analytical procedure to determine both ␦18O and ␦17O of H2O2 in natural water and first
measurements. Atmos. Environ. 33, 3683–3690.
Schemenauer R. S. and Cereceda P. (1992) The quality of fog water
collected for domestic and agricultural use in Chile. J. Appl. Meteor. 31, 275–290.
Searl A. and Rankin S. (1993) A preliminary petrographic study of the
Chilean nitrates. Geol. Mag. 130, 319 –333.
Seinfeld J. H. and Pandis S. N. (1998) Atmospheric Chemistry and
Physics: From Air Pollution to Climate Change. Wiley, New
York.
Sievering H., Lerner B., Slavich J., Anderson J., Posfai M. and Cainey
J. (1999) O3 oxidation of SO2 in sea-salt aerosol water: Size
distribution of non-sea-salt sulfate during the First Aerosol Characterization Experiment (ACE 1). J. Geophys. Res. 104, 21707–
21717.
Thiemens M. H. (1999) Mass-independent isotope effects in planetary
atmospheres and the early Solar system. Science 283, 341–345.
Thiemens M. H. (2001) Perspectives: Atmospheric science. The massindependent ozone isotope effect. Science 293, 226.
Thiemens M. H. and Heidenreich J. E. III.. (1983) The mass-independent fractionation of oxygen: a novel isotope effect and its possible
cosmochemical implications. Science 219, 1073–1075.
Thiemens M. H., Savarino J., Farquhar J. and Bao H. (2001) Mass-.
Independent isotopic compositions in terrestrial and extraterrestrial
solids and their applications. Acc. Chem. Res.34, 645– 652.
Thompson L. G., Davis M. E., Mosley-Thompson E., Sowers T. A.,
Henderson K. A., Zagorodnov V. S., Lin P. N., Mikhalenko V. N.,
Campen R. K., Bolzan J. F, et al. (1998) A 25,000-year tropical
climate history from Bolivian ice cores. Science 282, 1858 –
1864.
Thompson L. G., Mosley-Thompson E. and Henderson K. A. (2000)
Ice-core palaeoclimate records in tropical South America since the
Last Glacial Maximum. J. Quaternary Sci. 15, 377–394.
Vargas G., Ortlieb L. and Rutllant J. (2000) Historic mudflows in
Antofagasta, Chile, and their relationship to the El Nino/Southern
Oscillation events. Rev. Geol. Chile 27, 157–176.
Vong R. J., Baker B. M., Brechtel F. J., Collier R. T., Harris J. M.,
Kowalski A. S., McDonald N. C. and McInnes L. M. (1997) Ionic
and trace element composition of cloud water collected on the
Olympic Peninsula of Washington State. Atmos. Environ. 31,
1991–2001.
Wakshal E. and Nielsen H. (1982) Variations of ␦S34(SO4), ␦O18(H2O)
and Cl/SO4 ratio in rainwater over northern Israel, from the Mediterranean Coast to Jordan Rift-Valley and Golan Heights. Earth
Planet. Sci. Lett. 61, 272–282.
Warneck P. (2000) Chemistry of the Natural Atmosphere. Academic
Press, London.
Williams A. E. and Rodoni D. P. (1997) Regional isotope effects and
application to hydrologic investigations in southwestern California.
Water Resor. Res. 33, 1721–1729.
Wimschneider A. and Heumann K. G. (1995) Iodine Speciation in
Size-Fractionated Atmospheric Particles by Isotope-Dilution MassSpectrometry. Fresenius J. Anal. Chem. 353, 191–196.
Young E. D., Galy A., and Nagahara H. (2002) Kinetic and equilibrium
mass-dependent isotope fractionation laws in nature and their geochemical and cosmochemical significance. Geochim. Cosmochim.
Acta 66, 1095–1104.
Yung Y. L. and DeMore W. B. (1999) Photochemistry of Planetary
Atmospheres. Oxford University Press, London.
Yvon S. A., Plane J. M. C., Nien C.-F., Cooper D. J. and Saltzman E. S.
(1996) Interaction between nitrogen and sulfur cycles in the
4038
G. Michalski, J. K. Böhlke, and M. Thiemens
polluted marine boundary layer. J. Geophys. Res. 101,
1379 –1386.
APPENDIX
([OQO] ⫹ [OOQ]) ⁄ [OOO] ⫽ (k2n ⫹ k3n(Keq)⫺1)[OQ] ⁄ [OO]
Ozone ⌬17O values can be calculated knowing the rates of formation and destruction of the various isotopomers. The photolysis
rates for each isotopomer are assumed to be approximately equal
given their quantum nature and it is in the recombination process
that isotopic partitioning occurs. The three possible ozone isotopomer production channels are
O ⫹ OO ⫹ M → O3
k1
O ⫹ OQ ⫹ M → OQO x k2
→OOQ y k2
Q ⫹ OO ⫹ M → QOO
k3
Where Q is either O or O, x/y is the branching ratio for the OQ ⫹
O reaction which can form symmetric OQO (x) or asymmetric QOO (y)
ozone molecules, and k’s are the measured rate coefficients. The rates
of formation are then
17
Strobel 1983). The normalized rate ratios of k2n ⫽ k2/k1 and k3n ⫽
k3/k1 have been measured by Mauersberger (1999) for 18O (k2n ⫽ 1.27,
k3n ⫽ 0.93) and for 17O (k2n ⫽ 1.17, k3n ⫽ 1.03) giving
18
d ⁄ dt[OQO] ⫹ [QOO] ⫽ k2[OQ][O] ⫹ k3[OO][Q]
(17)
d ⁄ dt[OOO] ⫽ k1[OO][O]
(18)
In calculating the isotopic composition of the bulk ozone it is not
necessary to differentiate between the symmetric and asymmetric
branching ratio since x ⫹ y ⫽ 1, but this has been done by other
investigators interested in symmetric/asymmetric enrichments. Dividing (17) by (18), assuming steady state (d/dt ⫽ 0) gives
([OOQ] ⫹ [OQO]) ⁄ [OOO] ⫽ k2 ⁄ k1[OQ] ⁄ [OO] ⫹ k3 ⁄ k1[Q] ⁄ [O]
The isotopic exchange between O3(P) and O2 (Q ⫹ O2 ↔ OQ ⫹ O) is
many orders of magnitude faster than the O3 production rate and it
therefore determines the [Q]/[O] ratio. &rtf-symbol-84;he equilibrium
constant (Keq) for this exchange has been measured for Q ⫽ 18O
(1.94exp(32/T)) and calculated for Q ⫽ 17O (1.97(16/T)) (Kaye and
([OQO] ⫹ [OOQ]) ⁄ [OOO]
[QO] ⁄ [OO]
⫽
[QO] ⁄ [OO]O3
[QO] ⁄ [OO]O3
⫽
2
3
(k2n ⫹ k3n(Keq)⫺1)
⫽ 1.133 18O ⫽ 1.110
17
O
(19)
The left-hand side of (19) takes the form of R/R in the familiar
isotope delta formula (1) with product ozone as the sample and the
initial O2 as the reference. Statistically only 2/3 of the O2 molecules
generated from conversion of OQO ⫹ QOO (19, middle) will
contain Q, hence the balancing factor of 2/3 on the right-hand side
of (19). Using (1) and (2) gives ␦18O ⫽ 133‰, ␦17O ⫽ 110% and
⌬17O ⫽ 40.8‰.
The rate constants used above (kn2,kn3) were determined at low
pressure (200 Torr) relative to tropospheric conditions. The experimental data on the pressure and temperature dependence of ozone
⌬17O values is more applicable, and is useful in examining the
magnitude of ozone ⌬17O variations under tropospheric conditions.
A second order polynomial fit of the pressure dependency data from
Morton (1990) yields ␦18O ⫽ ⫺1.1385p2 ⫹ 2.3185p ⫹ 11.559 and
␦17O ⫽ ⫺0.8979p2 ⫹ 1.8069p ⫹ 10.3, where p is the log of pressure
in Torr. The temperature data (adjusted for isotopic exchange), also
from Morton et al. (1990), give linear fits relative to T (K): ␦18O ⫽
0.035T ⫹ 4 and ␦17O ⫽ 0.0243T ⫹ 3.7667. For surface pressure of
1 ATM (760 Torr) and temperature of 298K, ⌬17O values of 34.8‰
and 34.7‰ are generated respectively. Temperature variations with
altitude in the boundary layer (⬃3km) are typically ⫾10 K resulting
in a maximum ⬃1.5‰ ⌬17O fluctuation and a similar value arises
from the pressure variability. Since altitudinal changes in T and P
effect ⌬17O in opposite directions, and only daytime T (ozone
production) variations are considered, the P and T effect is on the
order of 1‰. Since nitrate deposition to the Atacama is an annual
average and integrated over thousands of years, the O3 and NO3⫺atm
⌬17O values can be considered constant at ⬃35‰ and 24‰, respectively.